Analytics user interface & data dimensions
  • 11 Jul 2024
  • 7 Minutes to read
  • Dark
    Light
  • PDF

Analytics user interface & data dimensions

  • Dark
    Light
  • PDF

Article summary

Pending release

This is pre-release documentation.

The analytics portal is available as a web interface which allows browsing, printing and downloading of podcast or stream access data.

See Analytics overview for details on the metrics and data delivered inside this portal.

Interface section

The main interface consists of several functional areas:

Screenshot 2024-06-05 at 14.00.59.png

Navigation (section 1,2)

  1. Enables navigation between different datasets. Switch between analytics for our stream, podcast, ad, recognition and other products. Only products availabe to the logged in user account will be shown.
  2. Select from different dimensions available for the selected dataset. Available dimensions in this section may differ based on dataset selected and the podcast or stream package of the account.

analytics.nav.png

Content selection (section 3,4)

3.1 Selects which content to include in the report. Some packages allow multiple items to be selected simultaneously, but only allows the same type of resource to be selected at a time - for example combing a podcast channel and radio stream in one report is not possible.
3.2 Control the reporting period. Several preset values are available and the ability to select custom time ranges. The amount of historic data available is limited by the podcast or stream package of the account.
4. The reporting timezone can be changed under the user profile. This allows browsing the analytics adjusted to your local timezone.

Content filters (section 5)

  1. Data filters can applied to the selected content. This allows filtering by:
    5.1 Customer content is always filtered by "Qualified" sessions only.
    5.2 Viewing only access from specific countries or regions.
    5.3 Viewing only access by specific devices or platforms.
    5.4 Viewing only access during specific dayparts.

Same filter types behave as OR statements, different types as AND filters:

Show content where (Country 1 OR Country 2) AND (Region 1 OR Region 2) AND (Devices) AND (Platforms)

Example: selecting Malta and Greece as country filters will show traffic from both these countries, but adding Scotland as region filter will prevent any data from being visible, as no traffic can be from (Malta OR Greece) AND (Scotland).

Selection content (section 6)

The main viewport area with widgets & tables showing all the available audio metrics for the filtered data. The contents of this area is largely determine by the dimension selected in the sidebar. (section 2)

Data dimensions

More than 20 different dimensions are available, each focussed on highlighting a specific aspect of the selected data.

Overview

Summary view which includes all the key metrics, a time-series chart and links to the most relevant sections.

podcast.analytics.overview.png

The top metrics on this page shows the percentage change compared to the previous period. This uses a mirror period of the selected reporting period, for example when viewing August data, this will be calculated against July.

Time series

  • The time series chart shows unique users (Users) and active sessions (AS) over time.
  • The resolution of this widget adapts to the selected reporting range. Shorter ranges use the minimum 1 HOUR resolution, while longer periods may increase to a 1 DAY resolution to ensure a consistent number of data points.
  • A vertical annotation is placed at the start of every Monday in the reporting range. This is labeled with M if there are less than 10 in the reporting period, or reverts to a simple vertical line for larger sets to improve readability.
  • The CSV download offers additional options to specific the resolution of the downloaded data.

Streams, Channels, Episodes

Depending on the current dataset, this section provides a break-down by stream, podcast show or podcast episode.

podcast.analytics.channels.png

Published episodes

Podcast only. Shows content accessed by podcast episode, but filtered to include only episodes that was also published inside the reporting period.

Behaviour

Exposes listening patterns related to when and how much audio was consumed.

podcast.analytics.behaviour.png

The Time of Week, Day of Week and Hour of Day charts are sensitive to the reporting period. If the selected range does not include a full week's worth of data some entries may be missing. If the reporting period includes an uneven number of days, this may also skew the results.

Podcast analytics displays a Listen through rate chart on this page, while stream analytics displays a Drop-off rate chart. This difference is due to podcast data consisting of episodes of finite but variable duration while streams are continuous 24/7 broadcasts.

Podcast analytics displays percentage buckets to cater for different audio durations indicating how much of the files have been accessed by users, while stream displays how long users listen to the stream in one session, divided into 5 minute buckets.

Podcast apps and even web playback typically pre-downloads a large section of the episode audio before playback, skewing the "Listen Through Rate" towards completion. This effect will have a higher impact on shorter episodes, where the entire file is more likely to be downloaded for playback. Any podcast app with "Offline listening" functionality will also skew this result as the entire file is downloaded for later consumption.

Day parts

Content access bucketed by separate parts of the day, a common reporting method for broadcasters.

podcast.analytics.dayparts.png

Day parts are based on the United Kingdom's Radio Joint Audience Research (RAJAR) schedule of day parts:

Day part nameDays of weekHours of day
Breakfast peakAll06:00 - 10:00
Mid morningAll10:00 - 13:00
AfternoonMon - Fri13:00 - 16:00
PM driveMon - Fri16:00 - 19:00
AfternoonSat, Sun13:00 - 19:00
EveningAll19:00 - 24:00
OvernightAll00:00 - 06:00

Daypart ranges are not user adjustable.

Geography

An overview of geographic access patterns with an interactive map and table links of the most relevant geographic data.

podcast.analytics.geography.png

The map provides three different display layers:

  • Country choropleth, rendered on the first 3 zoom levels, highlights countries based on their audience size.
  • Region choropleth, on the next 3 zoom levels, highlights top level country region based on audience size.
  • Location cells, displayed on all zoom levels, clusters audience sizes into hexagons on the map.

All map base layer and choropleth layers are generated from OpenStreetMap data.

Besides standard zooming and panning functionality, the map implements:

  • A "Full screen" mode that allows richer exploration,
  • A "Home" button to centre the map on the cell with the highest user concentration.

To reduce noise, all map layers exlude data points that contain only 1 user.

Countries

Content access grouped by country.

podcast.analytics.countries.png

Countries are tracked by their official ISO3166-1 country codes.

Regions

Content access grouped by top-level country subdivision.

podcast.analytics.regions.png

Uses the the official ISO3166-2 administrative boundaries. This data is not always available, and 40+ countries do not currently define any official sub-divisions.

Cities

Content access grouped by city or town name.

podcast.analytics.cities.png

Combined geography

Combines the country, region and city data into a single view.

podcast.analytics.geo.combined.png

ISP's

Groups content by the ISP or telecoms provider of the listener.

podcast.analytics.isps.png

ISP data can provide insights into user connectivity, for example identifying mobile vs landline providers. For customers negotiating zero-rating agreements it's also valuable to understand how much of their data is flowing over the relevant ISP.

Technology

An overview of the technology used by listeners to access the audio content.

podcast.analytics.technology.png

As best as possible, devices are assigned to these categories:

NameDescription
PhoneMobile phones
DesktopDesktop & laptop class devices
TabletHandheld tablets, like the iPad
AssistantSmart speakers and home assistants, like Alexa
TVSmart TVs and TV set-top boxes
VehicleSoftware running on in-vehicle dashboards
WatchSmart watches
Entertainment SystemRadio and hi-fi systems that can play online content
Media PlayerPortable media players, like the iPod
ConsoleGaming consoles, like the Playstation
Audio EquipmentStudio and broadcast equipment, like Barix receivers

Platforms

Groups content access by operating system.

podcast.analytics.platforms.png

Applications

Groups content by the name of the application.

podcast.analytics.applications.png

Combined technology

Combines the device, platform and application data into a single view.

podcast.analytics.tech.combined.png

Referer domains

Groups content access by the domain name where the web traffic originated.

podcast.analytics.referer.domains.png

Whenever content is loaded from a browser, it may include a Referer header that indicates where the listener originated from.
Referer headers are only used by web browsers. A large part of the audience is typically listed as Direct traffic, which is any other third-party application consuming the content like a podcast or stream mobile app or a voice assistant.
The referer is usually a full web page URL, but this view groups by domain.

The misspelling of Referer is intentional due to the HTTP specification.

Referers

Shows the full list of Referer string without grouping by domain and including all the URL contents available to identify individual pages driving traffic.


Compatibility

The analytics dashboard works in all recent versions of popular browsers, and have been tested in Chrome, Firefox and Safari.

The dashboard has been designed to work optimally on medium sized displays in landscape mode. This includes all tablet, laptop and desktop screen resolutions.

The dashboard is fully functional on mobile devices but some UI aspects may not function perfectly.

Dashboard authentication requires cross-domain cookies and will not work in Safari's "Private Mode" or other private browsing features which disables these cookies.