DEMOGRAFT

NCE Bundle

Marketing Bundle

Monetization Bundle

Changelog

FAQ

About

Used by industries:

TELCO
8.1 February 17, 2021 (back 21.3). 
User interface:
  • New location filter beta with local time support
  • Added handling “broken” audiences (no permission, the value that is in the filter does not exist in the database any more etc).
  • Added “paste values into filter”, including validation and cleanup of pasted values, handling the situation when many values are pasted/entered. This is supported for attributes that have less than 100 000 unique values in the database (does not depend on how many values are pasted). 
  • The first iteration of exposing new H3 geoindex data type in heatmap tab
  • LDB4 Map Algebra supports also multitable set-up
  • Added Help Export/Import GUI for platform managers
  • Fixed: Chart order if configuration contains sortValues attribute
Back-end:
  • New data type in Sparrow: geoindex. Currently H3 and S2 are supported in 1D breakdown API endpoint.
    H3: https://eng.uber.com/h3/, S2: https://s2geometry.io/ 
  • Ad hoc filter is significantly faster now. This is needed for temp ad hoc lists that will be the main method our own apps use as filters.
  • Targeter Sparrow+LDB and LDB4 can be deployed on separate Java virtual machines on the same hardware
  • Added external upload list API, currently supports Postal Codes and AD ID lists
  • Added external and internal list sharing API for uploaded lists
Small product improvements:
  • Login screen shows copyright date 2021
  • Applied security headers
    • sameorigin for X-Frame-Options header
    • Added Content-Security-Policy header

8.0 January 20, 2021 (back 21.1)
  • Completed the development of the audience parser
    • Deep nested audiences are supported now
    • Added opt-in info that propagates out from the nested audiences
  • Added external API for uploading and managing lists (/api/rw/1/external/uploadedlists/{type}/new etc.). Added DG_UPLOADLIST_API authority to determine if the function is available for particular users.
  • Map Explorer supports multiple tables (LDB4 back-end supported in great extent this already before, only heatmap API had to be enhanced). Some admin nuances remain for the next version.
  • Nested audience size is now calculated from the column that is used internally for the join (not from the token primary key column). In the normal system state, these are equal, but the new behavior helps to detect inconsistency quickly.
  • Added Sparrow table metainfo support also to internal API (existed in external API already long ago).
  • LDB4 supports an array of S2 “pixels” as filter criteria (this means “give information about this particular S2 pixel” is supported now”. Added support for native hex form of S2.

7.11 December 15, 2020 (back 20.13.1).
  •  User admin app: New Targeter 2 app can be assigned to users in Partner management page (and related authority to the database).
​    LDB4/Map Explorer:
  • New feature: Classifier unique on maps, table and crosstable.
  • New feature: mini-table for unique aggregation in tables to get the list that forms the unique.
  • Filter dropdowns work differently: they do not show full content of LTF files, but only what exists in the database.
  • Table and Crosstable provide to user an option to hide all NULL values. This is not a query “NOT NULL”, it is just hiding items where NULL was present. The feature also changes how unique is calculated (with or without NULL). 
  • Integrated with the newest file manager.

7.10 November 26, 2020 (back 20.12).
  • New file manager, with search mode and file type filters that can be configured per app and use-case. New UX design.
  • Created fundamentally new “file access” method in the platform and exposed this for POI-s (not for workspaces, uploaded data yet). Triggered the need to change also the way existing feature “Publish global POI layers” works. Added new authority DG_INTERNAL to keep internal and external Partners separate in the POI publish context.
  • Support Sparrow first level NOT, needed for NOT in nested audiences (NB! Works differently from NOT inside Sparrow attributes).
  • Metainfo API endpoints are now faster, dropdowns in filters populate more quickly.
  • Hamsterwheel is now able to start successfully even when translations are not in sync with the database structure. 
  • Added Sparrow reload multinode support and also support loading new attributes (useful for testing).
  • Security patch for RW servers

7.9 October 27, 2020
  • Significant facelift and layout change of Targeter
  • Included newest File Manager (3.0.9) to Targeter
 
7.8 October 1, 2020
  • Added “sparrow only” for the Nested Audience feature
  • Upgraded File Manager in all products (triggered by the Nested Audience feature)
  • In Targeter, improved the speed of export XLS tab
  • If non-existent new ad hoc is used in filter, respond with decent error to make situation handling in Javascript easier

7.7 September 15, 2020
  • Improved filtering by long-long data type
  • Added “int codec” to support more effectively the Nested Audience feature
  • Improved query speed while using EpisodeDB frequency filter
  • Export of id-s using AMQ from long-long and classifier fields is now more effective (not via Rest API)
  • Fixed bug related to “remember me” during sign-in
 
7.6 August 11, 2020
  • Completed “upload any data” framework, comes with Postal Code and AD ID parsers (other old parsers for MAC ID, MSISDN still need to be added). Few tweaks remain to make usage by Javascript easier. Not exposed in GUI yet.
  • Added support for suspending inactive accounts. Changed the main menu item from Password Policy to Account Policy.
  • Enhanced POI upload API to support more generic map content (also polygons). 
  • Fixed: Map layers modal POI layers checkbox disappears in Firefox
  • Fixed: Help icon tooltips are behind the help modal
  • Fixed Targeter info window that did not show versioned attributes correctly
  • Dwelling type chart height fix in Targeter
  • Help bugfix - images were doubled when pasting
  • Various small improvements and fixes in LDB4 Map Explorer

7.5 July 21, 2020
  • Introduced EpisodeDB Input frequency filter, AND/OR/NOT support. Not exposed in GUI yet 
  • Long worker supports re-try concept (e.g. should some calculation fail inside attribution formula, e.g. server restart, it is possible to ask Longworker to try again)
 
7.4 June 16, 2020
  • Added long-long 128 bit list data type
  • Support direct vector tiles without style.json from another host
  • ActiveMQ client version upgrade to 5.15 in Restfulweb, Datamanager and Gatekeeper
  • Capability to change Sparrow translations without Hamsterwheel restart
 
7.3 May 20, 2020 (back 20.4)
  • Help text and help images are stored separately in Mongo, this makes help migration between servers much easier
  • In Location tab, do not run tourism and stopLDB queries before modals are opened any more

7.2 April 22, 2020 (back 20.3)
  • Introduced fundamentally new module “longWorker”.
    This is designed to run scheduled/queued calculations that
    • Take longer than 1-2 minutes and end-user does not expect the result to be shown “right after the loader has stopped spinning”
    • Need to access databases other than LDB/LDB4
    • Deployable separately, without upgrade/restart of the rest of the platform
·       Changed the code of Targeter to fit long (French) texts
  • Fixed “Last week sometimes chooses current week instead of last available full week”
  • Fixed “Behviour filters don't apply after changing time for geofence”
  • Updated Map Explorer Pro to support Covid-19 analysis
  • Map Explorer help files are separated from each other better
  • DOUBLE field type support in Map Explorer
 
Hotfixes to 7.1.1 between March 13-25, 2020
  • AND-OR-NOT filter bug - sometimes toggling AND-OR-NOT switch did not have an effect on the filter and filtering was done by OR
  • Bug of Targeter NOT filter not showing up properly in audience info window fix
  • Audience trend fix
  • Resolved some issues with smaller screens in Data Manager
  • Fixes to Mongo driver in Query Engine
  • Fixed LDB4 Histogram aggregator ArrayIndexOutOfBoundsException
 
7.1.1 March 11, 2020
  • AND/OR/NOT improvements in Targeter
  • Behaviour filter disappeared in some cases. Fixed now.
  • Released new Map Explorer Pro
  • DOUBLE datatype support in LDB4 database engine
 
7.1 March 5, 2020 (back 20.1)
  • Added support for NOT for Sparrow attribute filters (backend)
  • Added three different dark Google Map styles
  • Newer version of File Manager (3.0.0) 
    • Handles now better multi-megabyte uploaded files, supports glTF and other smaller changes)
    • Improved graphical design
    • User-uploaded POI layers can be renamed by owner of the layer
    • Other small changes
  • Added general editable online help to Gatekeeper (Admin App)
  • Upgraded AMQ to 5.15.11 in Hamsterwheel (elsewhere will be done later) 
  • Restful Kibana proxy improvement (the functionality remains the same), prerequisite for upgrade to ELK 7.5.2
  • Small UX improvements in Gatekeeper
  • Slowness of Compare Behavior tab locked the application, now loading happens in the background
  • Removing authority from user removes now also authority from API key
  • UTF-8 French characters broke audience import-export, fixed now
 
7.0 January 8, 2020
  • Upgraded Spring from 4.1.6.RELEASE to 4.2.5.RELEASE
  • POI layer management API endpoint is now properly back-end permission-based
  • Data Manager user can unshare also POI layers that other Data Manager user globally shared
  • Removed top20 limit from tier1 interests chart in behaviour tab
  • Removed double x in filters if the filter text is long enough
  • Fixed export of ID-s (when it is configured to use Ad IDs and token had only one Ad ID)
 
6.8 December 4, 2019
  • 20+ million token support
  • 2+ GB file support for Sparrow attributes (for LDB was delivered already earlier) 
  • Ironed out some corner cases of API key feature
  • Improved error handling in KML/CSV export v2 in Data Manager
  • Couple of small bugs that are related to online help
  • Netscout driver writes data to Elasticsearch per time bucket, not cumulatively
  • User Admin GUI: improved wording in password policy
  • User Admin GUI: added user authorities info window to user list
  • The whole platform can be turned into read-only mode (e.g. Data Manager cannot change help, various settings and other things that are stored as public Mongo)
  • Timeline diagram and time filter had a bug that appeared when the Daylight Savings period ended
  • Solved so-called “Production Kibana login problem”: Gatekeeper session id was shared between domains, change it to subdomain level
  • Login spinner kept spinning when logging in with a suspended user
 
6.7 November 7, 2019
  • Added the concept of Global POI layers
  • Added SHAPE column to POI layer
  • Long title fits better into POI info box
  • Editing Streetview 15+ times did not work, now works.
  • Improved the handling after POI search - user uploaded POI layer symbols were not clickable and RADIUS/SHAPE columns were hard to use.
  • POI layer download works now properly even when user has not provided STREETVIEW_URL column in uploaded file
  • Added settings to toggle POI layers between predefined symbols and larger dots.
  • Removed all but 'Cells' subtabs from DataManager Statistics tab. We are using now Kibana for this data and removed old legacy code.
  • When relative time is used and time is out of available time range, GUI shows notification. This is useful in situations when deploy did not happen on-time for some reason.
 
6.6 October 2, 2019
  • File manager is now out of beta status.
  • “Reduced the memory footprint of LDB” in release 6.5 reduced the memory footprint, but this made restart time slower. Now we fixed the slow restart, it is back to where it used to be
  • Click on POI works now also inside lasso and geofence. NB! In edit mode, objects inside geofence remain not accessible.
  • It is now possible to create Google Streetview also to objects that did not have Streetview data in uploaded CSV (previously it was possible to edit, now user can also create). 
  • It is now possible to paste into Google Dynamic Streetview window view settings (previously only copy was possible). This is useful for situations where Streetview uploaded location had no easy way for navigating out (e.g. inside shop, photo sphere). 
  • Final iteration of version picker in Behavior tab (a.k.a. remove timestrip), comments on previous releases are taken into account.
  • Logs have now time and size-based rollover
    • Query engine
    • TV query engine
    • Restfulweb
    • Data manager
    • Gatekeeper 
  • LDB4: All queries are faster in single user scenario (multi-core support)
 
6.5.1 September 12, 6.5.2 September 19, 2019
  • Deployed newer home algorithm that supports “excluded areas” (e.g. airports, industrial areas).
  • Added support to store more documents to Elasticsearch index, useful for storing data about processes that do not run continuously, but for example daily. NB! This is only support, ETL does not take advantage of it yet.
  • Fixed empty heatmap and double scroll in timeline bugs
  • Fixed Top 500 and Top index modal closing X misalignment on hover
  • Fixed “Targeter dropdown should open with mouseover”
  • Fixed Resize chart size in Targeter
  • Fixed “Population pyramid height doubles while loading”
 
6.5 September 2, 2019
  • The first version of user-uploadable POI feature
  • The final version of incoming tourism feature
  • Added Kosovo into travel destination lookup table and inbound tourist lookup table
  • KML next-gen export API call is ready, waiting for Javascript availability to expose in GUI
  • Reduced the memory footprint of LDB (that was prominent for 100 000+ strong celldata)
  • Added “Data changed here” notification to timeline
  • Enable HD mode sooner in Targeter: 1850px -> 1500px
  • ETL can add more files to the same index (necessary for upcoming ELK 7 upgrade)
  • Data Manager: Apply Marketshare button did not work
  • Fixed bug when editing campaign with custom filter
  • Improved messages when too heavy LDB4 queries were executed
  • All LDB4 queries that use aggregation over more than one field are faster in most cases (multi-core support)

6.4.1 July 15, 2019
  • The first version of tourism airplane icon
  • First time ever delivered LDB4 visualisation tool Map Algebra (for demonstration purposes)
  • The first version of KML next-gen export (for demonstration purposes)
  • More changes to support better small screens and low resolution
  • MongoDB Java driver update
  • LDB4: Improved the speed of classifier filter for high cardinality fields (makes Tourism feature faster for large geofences).
  • LDB4: Fixed heatmap bug that appeared for negative S2 bins.
  • LDB4: Replaced “band aid” Boolean field handling with the proper solution

6.4 July 9, 2019
  • Phase 1 of increasing token max limit from 20M to 40M to analyze the impact on average query speed.
  • 2+ GB file support for Hamsterwheel
  • ELK 7.0 support in backend
  • Long list of online editable help improvements (UX, handling of WAF blocking save sometimes, etc.)
  • Added online editable help also to Data Manager
  • Various changes to support better small screens and low resolution
  • LDB maker can filter now events by cellType (will be used in the future to remove 3G events from Netscout) and MCC to avoid injecting outbound roaming data into data
  • Daily report various LDB size statistics to Elasticsearch
  • Drake Java protocol permits sending also tags to Elasticsearch
  • Sparrow can now ingest new data without daily restart
  • Improved KML visualisation - tooltip is available also for lines, not only points. Delivered API endpoint for next-gen KML visualisation also, for testing, it is not available via Data Manager GUI yet.
  • Reduced the timeout of httpclient connections, this probably caused application freeze after 6.2 upgrade.
  • ETL refuses now creating larger LDB files than 2GB. 
  • LDB4: Made max unique value limit in classifier field configurable. The default is 10 000, but 500 000 should also work.
  • Map Explorer GUI
    • Added locationType to filters
    • Added App Launcher
    • Improved slider handling on heatmap color bar
    • Improved heatmap bin info panel
  • Fixed: Interest global trend had “New data source” outside of diagram
  • Fixed: Targeter Heatmap went empty (color string was broken) when sliders were moved very close to each other
  • Fixed: Audience size went 0 whey user clicked RUN too fast
  • Fixed: Two bottom diagrams in Timeline tab were under each other, not beside each other
  • Fixed: Editing audience that had classifier values outside of first 100 was broken (so-called FSA filter bug). 

6.3 May 7, 2019
  • LDB4: minor polishing of logging, better handling of oversize breakdowns
  • LDB4: fixed issue of requesting heatmap detailed information on Western Hemisphere
  • Numerous small UX changes (“Display "selected version" actual version in bold”, “Reduce the X size in the info window”, “Display segment size number in bold”, “new icons in audience list”)
  • Show top/bottom 500 in top index modal (previously top/bottom 250)
  • Completed leftovers from of versioned binary picker development
    • support for audience validation
    • Finalized ETL using Javascript for parsing audience definitions
  • Help improvements
    • Added source code view. Helps removing hidden characters that can happen when text is pasted from Word
    • Added 16 and 20 font sizes to Help editing options
  • Cleaned up discontinued permissions in Gatekeeper
  • Fixed bugs:
  • that made extra getBreakdowns query on heatmap that was never used
  • Edit mode in heatmap tab created ad hoc dimensions that are not used any more.
  • Breakdowns time filter disappears when changing geofence size
  • Broken audiences did not work as designed after upgrade to 6.2.

6.2 April 17, 2019
  • The first iteration of versioned attribute picker feature in filters, Analyze and LDB4: added IS NULL and IS NOT NULL filters for classifier fields
  • LDB4: added a new timing log which enables analysis of database performance
  • LDB4: geofence is not mandatory any more for aggregate API endpoint
  • Implemented “Query String Parameter in HTTPS Request”, “Clickjacking - JS <style>html”, “directory listing enabled”, “direct access to admin pages (geoserver)”, “Vulnerable software version”/”Do not disclose Kibana version in response headers” vulnerability scan findings and “Unauthenticated file access (api/rw/1/conf/shared)” 
  • Changed text that is displayed above timestrip in Behaviour tab
  • Increased the number of custom filters per partner from 100 to 200
  • Improved handling of the situation when user wants to access Kibana without logging in to the platform first
  • Removed "Opt-in size" in Order campaign dialog
  • Final tweak of UX facelift (font sizes etc)

6.1 March 25, 2019
  • LDB4: Refactored method of deleting the data
  • LDB4: Times out gracefully very difficult queries and over-limit classifier data
  • LDB4: Statistics aggregator, AND filter for LIST type, automatic fallback for lower resolution heatmaps
  • LDB4: Added UNIQUE aggregator
  • LDB4: refactored support for LONG and CLASSIFIER data types
  • User management: “the last man standing” admin user cannot delete himself/herself
  • Handle the situation when audience geofence is in ancient past, outside of last 13 months of Behavioral time strip
  • Heatmap modal improvements (is open now again by default, minimizes and does not close)
  • Fix the issue caused by Google Maps API change (controls and base map selector grew big)
  • Filter panel header shows now how many filters are applied for current audience)
  • Various UX fixes

6.0 (February 5, 2019)
Main changes:
  • Test/dev GUI feature of multi-LDB geofencing (a.k.a. Trade area), available only for explorer users. Back-end was delivered already earlier. Currently comes with only few weeks of stop-LDB files, for testing.
  • Added MAC id upload feature with associated user permission, changed “Partner CRM Upload” to more generic “Upload data”. Improved the existing functionality of handling erroneous lines - now we show them all at once.
  • Added user-selectable breakdown widget to Heatmap tab (this is a copy of the feature we had in Compare Heatmap already).
Other changes:
  • User can enter wrong password now 5 times (was 10 times) before leaky bucket throttling kicks in
  • Order Campaign dialog is now much more verbose
  • CMD+Enter on MacOs now runs query like CTRL+Enter does in Windows
  • Filter panel group collapse icons are now behind the title
  • Login app no longer displays French EULA to new user
  • Location data importer can now handle “ inside text fields
  • Netscout driver can now gracefully skip non-numeric data in numeric fields

5.8.1 (November 23, 2018)
Main changes:
  • 1:N relation for TV-Mobile join
Other changes:
  • UX improvements in Attribution tab
  • Changed the lift formula in Attribution tab
  • Ordered broken campaign audiences are now visible to campaign manager
  • Fixed too many decimals in Population pyramid, Google map control buttons, Ctrl+Enter behavior in Targeter

5.7 (November 1, 2018)
Main changes:
  • Individual login screens of apps replaced with unified login app
  • New heatmap H algorithm. Cell density heatmap uses new H, all other heatmap types use old H algorithm. This is OK, as the issue was in cell density map, not in other heatmaps.

5.6 (October 10, 2018)
Main changes:
  • Added multi-tenancy support for MongoDB in Restful servers, one HW can support several RW instances.
 Small changes:
  • Using ENTER in filter panel works better now
  • CTRL+ENTER runs query now both on TV and Mobile app
  • Removed hairline from diagrams, added search icon to address search
 Small bugs:
  • Compare Heatmap and Compare Timeline and Footfall tab corner cases 

5.5 (September 25, 2018)
Main changes:
  • MultiLDB support first phase, ready for testing ‘in vivo’. Only getBreakdowns request supports it, others need to be added, if testing goes well.
  • Added support for multiple browsers in “remember my 2FA” to make functionality more similar to “remember my computer”
  • Possibility to manage users also without SMTP integration
  • Possibility to send start-end times and errors of any (scheduled) process from any server to central Elastic.
    • Drake needs to be updated and configured also to take advantage of this. 
 Small changes:
  • Geofence management icons are now better aligned with Google Maps icons (Google Maps triggered the need for this change)
 Small bugs:
  • Geofence trash icon disappeared sometimes

5.4 (Early September, 2018)
Main changes:
  • Added population statistics support to XLS export functionality
Small changes:
  • Removed wicket wkt widget from Javascript (3PP license issue)
  • Reduced clutter in KML export (does not dump empty geoLDB layers)
  • Increased the Partner CRM limit to 200 lists
  • Bulk audience export tool can export many audiences in .zip file  
 Small bugs:
  • Various compare heatmap and audience save corner cases are fixed

5.3 (May 11, 2018)
  • 2FA using TOTP (time-based one-time password):
    • OD matrix and TV do not support it yet
    • End-user voluntarily setting up 2FA is not included, this is in the roadmap
    • “remember me” feature not included, this is in the roadmap
  • French translations updated.
  • ?-button on diagrams in Industry and Multicompare tabs tell now what version was used (for versioned attributes) to make them behave similarly as in Audience and Compare.
  • Bulk tool - prohibited importing multigeofence audiences when number of geofences is > 10; prohibit importing when audience count will be larger than 1000 after import.
  • Animate (beta) tab deprecated, not available any more.
  • LIST type is now much faster when running queries where a lot of items is in filter (this is the case for “Chinese websites” audience, for example, where a lot of websites are picked for the filter).
  • Reduced unnecessary queries and activity when saving audience and returning to the audience list.
​Bugfixes:
  • Heatmap location was not handled correctly.
  • Tooltips did not disappear sometimes.
  • XLS tab: De-selecting dimension from being exported did not re-locate it into correct location alphabetically sometimes.

5.2.1 (April 12, 2018)
  • New “FB-style” loaders are now also on Compare and Multicompare.
  • Heatmap controls has now three non-cell heatmaps: FSA, CSD, CSDADA (actually much more, but hidden behind thunderbolt symbol).
  • Switched from old FSA borders to FSA2016. This affects all – heatmaps, breakdowns, filters, footfall. For this, edited 7 FSA->City mappings.
  • On Heatmap: instead of old “borders on/off” user can select what borders to show (FSA, CSD, CSDADA).
  • Audience bulk tool supports also JSON files that are in ZIP container (user can import many audiences together). Effectively this is a mirror of Export ZIP feature.
  • Re-wrote bulk tool geocode from server into user browser (because of firewall restrictions).
Small feature improvements/changes:
  • Last login report in user manager works now differently. In the past the timestamp was “last time user tried to log in (but maybe was denied because of LDAP”. Now the timestamp shows when the user really was let in. This means all LDAP-denied users
  • Compare heatmap has settings wheel button to hide all mini-windows on screen (not collapse, but hide completely). Heatmap controls mini window is also now freely movable. Added tooltips to view mode.
  • Custom time heatmap is not possible on non-cell heatmaps any more, only on cell heatmap.
  • Import audiences: selecting color takes automatically to file selection (removes the need for extra mouse click).
  • Changed texts in Data Manager Marketshare tab to explain the feature better.
  • Button hover style while focused is now blue instead of grey.
  • Sankey diagram is permanently removed from Mobile tab.
  • Reducing flicker: When saving audience, total number was animated two times, when moving orange time in behaviour, now we only animate trend column.
  • Removed option to edit Python code in Data Manager GUI. HTML is still editable. We do this in command line over VPN and never used this feature.
 Fixed bugs:
  • Heatmap opacity slider did not round % and erased heatmap when moved from 0 to 100 quickly.
  • Saving heatmap saved wrong map window location, compare heatmap did not use correct map window location.
  • Compare heatmap did not show heatmap type correctly and had FSA borders under heatmap.

5.2 (March 22, 2018)
Main changes:
  • Fully integrated Bulk Audience Generation Tool. Input file format is now XLS (not CSV). Tool supports three methods that user can freely mix: lat+lon+radius, address geocode, wkt. Added preview step to avoid importing erroneous/not permitted audience definitions. Tool has downloadable sample file and help, both can be edited in Data Manager.
  • Modern Facebook-style loaders v1. To enable this, we re-wrote all Kendo bar diagrams from scratch in HTML. Added fade in-out and animated transition that works differently for TOP and fixed order diagrams.
  • Imported new data (2016) of borders for heatmaps. FSA borders have changed a lot, we added also CSD and experimentally CSD+ADA, DA and other borders.
  • Permission-driven subsets of rows. The primary usecase is giving access to “blacklist” or not, but the feature is more generic.
  • Permission-driven subset of columns v1. The primary usecase currently is to avoid exposing household and some other attributes to external users. Added special permission to enable/disable dimensions that are not part of any tab. Version 1 hides dimensions from GUI, not yet from API.
  • The platform has now two partner types and EULA-s: internal and external.
  • Add possibility to freeze number of subscribers to avoid exposing growth before it is made public.
 Smaller changes and bugfixes:
  • Implemented numerous small color and layout changes.
  • Versioned filters in Industry were not supported, now they are.
  • Changed household income formula and added Prizm5 to Household tab.
  • Fixed custom filter dimension handling when user had not saved the change.
  • Improved handling of audience that attempts to show missing data on Timeline (happens when data feed is frozen).
  • When audience was in 2017, behavioural yearly view did not work properly.
  • Changed global trend text into blue to follow color coding.
 Product improvements:
  • Completely re-written “secondary” heatmap feature. They are now 10-50x faster, and support many, not only one secondary heatmap. QE CPU load is greatly reduced, Collector layer needs to work a bit more, and RAM requirement goes somewhat up on RW nodes (depends on how many secondary heatmaps we will keep).
  • Added overlap and breakdown widget to Compare Heatmap. User does not need to switch tabs to quantify heatmaps.
  • Added side-by-side and detached views to Compare Heatmap.
  • Upgraded ELK stack. This is major change, requires conversion of log database content to new Elasticsearch format. Kibana GUI is also much different (and better).
  • Improved the design on the dialog of management of custom filters (feature was delivered in 5.1.1, now added only design).

5.1.1 (February 6, 2018)
Main changes:
  • LDAP support improvement: User Admin GUI has sortable column that shows the date when user successfully logged in last time.
  • Export Excel tab handles better long list of dimensions and has configurable max number of breakdowns in XLS export request to avoid overloading the system.
  • Implemented the initial proposal of Facebook-type of loader on Behavior tab.
  • Ad hoc management panel can now delete all and supports two deletion modes: ad hoc only or ad hoc with associated audiences. Graphics design will follow in next version.
  • When 3PP API developer composed ill-formed CRM filter, we did not return an error, but returned all users. This was misleading. Now the response includes error.
  • Modernized Mobile tab (all diagrams are in HTML and use breakdown+index approach as other tabs). Phone model names are made nice with additional lookup table.
Smaller bugs/things in GUI:
  • Display message in the MM login screen that IE is not supported.
  • Global trend - reduce height of the versioned block displayed under x axis.
  • 2017 must start at the start of Jan not in the middle.
  • Re-order "Credit score" in Household tab.
  • Global trend must not show 2018 data in 2017.
  • Breakdown bar height was off whereas index was OK.

5.1 (January, 2018)
Mobile App:
  • Added Household tab.
  • Finalized versioned Travel destinations (in Audience, Compare, Multi-Compare mode, in filter panel).
  • Third iteration of Global Trend (normalization, legend, blue and orange bars etc).
  • New user-selectable setting "Charts always in HD".
  • Behavior timestrip supports now three years (needed in the beginning of the new year, as we need to display 13 months+rolling 30d).
  • Improved PDF Help: Data Manager can delete PDF and end-user does not see message “PDF help not available”.
  • Improved context-sensitive help behavior in edit mode (follows now tabs that user has selected).
  • When user A makes a copy from audience that user B has shared and if the audience includes custom filters, then these custom filters will be copied also into user A collection. This situation is shown also in edit mode and custom filter edit window.
  • i-window shows version for every versioned dimension that is used in the filter (because each dimension can have different versions available).
  • Non-Explorer users cannot create query that exceeds configured amount of cell*hours (currently set to 4 million). NB! only in Edit mode, API and bulk audience tool users can still bypass this limit.
  • User can add geofences with POI search without hitting Save after each save.
  • Back-end restart is now faster (huge number of versioned list type dimensions slowed the old code down), saving about 15 minutes from daily deployment time.
  • Back-end node is able to tell to cluster restart scripts that restart of node is done and the next node can proceed with rolling restart. This is not configured yet. Will save about 15 minutes from daily deployment time.
  • Location tab POI/address Google search box is now omnibox and lets user to enter latitude,longitude.
  • Top Index modal applies now size as secondary sort (needed when many horizontal bars have the same index value).
  • Footfall tab calculates radiuses better for audiences that have few people and small geofence combination.
  • All versioned (including TV) filters are now fully supported in Data Manager.
  • User who can do only opt-in export cannot export if audience includes less than L users anymore.
  • Upgraded embedded component Geoserver to 2.11.
 Bugfixes:
  • Compare and Multicompare A vs B uses now correct version in Behavior tab.
  • When user navigates back from edit mode immediately after Save, confirmation message “do you want to save?” should not appear again.

5.0.3 (December 19, 2017)
Main changes:
  • Fix “700+ audience” issue. This version works with 1000 quite OK, but if we want to support 2000+, then we need bigger redesign of the component.
  • The first version of Versioned travel destinations (diagram in Places tab).
  • Show three trend numbers in global trend.
  • Added business/consumer segment filter.
 Small feature improvements:
  • Report manager has new and improved GUI (and quite important change (“sandbox”) behind the scenes).
  • Javascript triggered unnecessary getBreakdown queries, we removed these.

2.9 (July 2, 2015)
  • GUI for a platform manager for changing the list and order of filters available for regular users (previously was available for technical support via configuration file)
  • Attributes of cells migrated to the same database engine where data of subscribers is stored. This will harmonize the ETL process and handling saved queries

2.8 (June 3, 2015)
  • New Query: Breakdown by distance for geotags. Respective diagrams and maps in Targeter
  • New platform feature: per-user ad hoc attributes. Users can save result of every query as a new attribute in the subscriber database, and used as input filter for the next query
  • Faster server-side method for pre-completion of filter values. Makes user interfaces much faster

2.7 (April 30, 2015)
  • New data type: spatial (aka geotags). Database of subscribers now supports locations expressed as geographic points, these attributes can be used for filtering (e.g. “work”/”home” inside/outside given radius)
  • New query: overlap unique IDs of two query results
  • New Targeter feature: Compare query results

2.6 (March 30, 2015)
  • Improved mechanism for saving query definitions, supporting sharing with colleagues and favourites
  • Subscriber database now supports end-user uploadable attributes (lists). Useful for third parties for uploading e.g. their loyalty card members
  • Platform-configurable seed query that will be created for every new user to help them to get started faster
  • Concept of relative time in the query definition (e.g. “yesterday” instead of given date can be used)
  • Completely rewritten Targeter application

2.45 (January 20, 2015)
  • Swapped the web framework to Spring
  • Database engine that serves statistics will be much more flexible and powerful (Mongo)
  • Authentication and authorization module rewritten, new user interface for platform user and role management. Dedicated role for password policy enforcement
  • Support for value-based pricing (i.e. provisioning features by user account). Useful for differentiating the platform usage fee

2.3 (November 4, 2014)
  • Numerous fixes and improvements before upcoming major upgrade

2.27 (October 21, 2014)
  • Server-side map layers make the display of large amounts of POIs much faster and support several thematic map layers
  • Updated KendoUI v2014.2.903, the library that we use to display graphs and most of other user interface components

2.23 (September 22, 2014)
  • NCE: When selecting particular usecase, appropriate map layers are turned on automatically.
  • Long list of usability fixes in NCE and Targeter
  • Better support for location data of variable accuracy (some events can be processed at Cell ID accuracy, other events at SAC/site accuracy)
  • Statistics: URL now features hashtags for easier navigation and bookmarking ( e.g. /statistics#cells)
  • Statistics: new charts for average number of cells and average number of events per subscriber to evaluate the data quality
  • Axis on charts use k, M and B (for thousands, millions and billions respectively) to make labels on charts shorter

2.1 (August 7, 2014)
  • Each user can have own POI list
  • Support for satellite maps
  • New audience index report (compare audience to the base per attribute) in Targeter
  • New (d)CRM field type: list. For regular attributes one MSISDN can have only one value (e.g. gender can be either ‘male’ or ‘female’, not both), list type enables storing many values (e.g. interests in ‘football’, ‘windsurfing’ and ‘gadgets’

2.0 (June 4, 2014)
  • Completely rewritten RDBMS support for managing cell data
  • New statistics
  • New version of the map component. Much more modular, supports data-driven GUI. Supports now also search from POI list and saved geofences
  • New: concept of bookmarks that can be saved in Demograft database, corresponding changes in API
  • Much more robust Collector module to merge multiple CDR imports from different formats
  • “Heatmap of money” added to NCE
  • Improvements to make deploy process faster and less error-prone

1.60 (Apr 22, 2014)
  • Added Top 20 sites animation to NCE (requires newer browsers)
  • Fixes and code optimizations

1.53 (Mar 27, 2014)
  • Completely new version of NCE: supports more usecases in GUI, adds redesigned filters for cell KPIs, adds respective set of queries in API and much more.
  • Statistics in beta. Statistics gives a helicopter view on the data in Demograft. The goal is make sure that we will not have garbage in – garbage out analytics platform.
  • Much improved logging of the platform to improve testing and maintenance efficiency
  • Updated Visual Network

1.50 (Feb 12, 2014)
  • NEW feature: time intervals. All queries can be executed e.g. for Monday afternoons (Note: heatmap does not support this yet)
  • NEW in Targeter: grab from any filter title and drag it to breakdowns area to see the diagram by this attribute
  • Automatic cell data processing algorithm is improved
  • Long list with smaller Targeter GUI finetunings

1.40 (Jan 17, 2014)
  • Most of queries are now way faster, due to new ldb index
  • Targeter: moved several GUI elements to advanced more to give less cluttered view for regular users.
  • Targeter: numerous small fine tunings, most of them stemmed from major changes introduced with ver 1.30.
  • Targeter: we returned to our initial design – show geofencing results and heatmap buttons are again visible on beside each other. Enjoy!

1.39 (Dec 13, 2013)
  • Use /developers to learn how Demograft API is working (beta)
  • /visualnetwork out of beta and ready
  • Numerous bug fixes
  • Last but not least – Merry Xmas!

1.31 (Nov 8, 2013)
  • Targeter shows cells that are used in geofencing and the Targeter queries can be easily copied into bulk query tool definition file

1.30 (Oct 25, 2013)
  • Completely rewritten Targeter application GUI
  • Visual Network now in beta status

1.27 (Oct 04, 2013)
  • Cellplan import mechanism and algorithm updated

1.22 (Sept 18, 2013)
  • NCE has several usability and configuration improvements
  • Bulk query tool has several usability improvements

1.21 (Sept 13, 2013)
  • Fixed four bugs (heatmap is not displayed by default anymore, map not loading sometimes after application initialisation, Targeter time selection has 0:00 hours as default irrespective to the timezone setting, heatmap now honours privacy K-limit as Demograft queries do)
  • Added bulk query tool as pre-alpha version.
  • Added network visualisation tool as pre-alpha.

1.15 (August 27, 2013)
  • Added Heatmap in beta status
  • Simplified navigation and streamlined GUI of heatmap
  • Added histogram to the heatmap. Histogram shows the territory covered with particular color (=density of people).
  • More and nicer color schemes for heatmap
  • The direction of heatmap colors is reversed – maximum is at right now
  • Server side logging to measure the performance of heatmapping
  • Heatmap has two modes: “trails=true” is default and counts person events (if someone was detected in given cell 10 times during given time period, it counts as 10). “trails=false” counts unique MSISDN-s, no matter how many times they were observed in given cell.
  • Experimental algorithm, not polished yet: “relative=true” gives you ratio of selected filtered people into the total number of people observed in given area. E.g. Heatmap does not show absolute number of iPhone users, but share of iPhone users into all phone users.

​Privacy Policy
​© Reach-U 2022