What is the MAT?

OONI's Measurement Aggregation Toolkit (MAT) is a tool that enables you to generate your own custom charts based on aggregate views of real-time OONI data collected from around the world.

OONI data consists of network measurements collected by OONI Probe users around the world.

These measurements contain information about various types of internet censorship, such as the blocking of websites and apps around the world.

Who is the MAT for?

The MAT was built for researchers, journalists, and human rights defenders interested in examining internet censorship around the world.

Why use the MAT?

When examining cases of internet censorship, it's important to look at many measurements at once ("in aggregate") in order to answer key questions like the following:

• Does the testing of a service (e.g. Facebook) present signs of blocking every time that it is tested in a country? This can be helpful for ruling out false positives.
• What types of websites (e.g. human rights websites) are blocked in each country?
• In which countries is a specific website (e.g. bbc.com) blocked?
• How does the blocking of different apps (e.g. WhatsApp or Telegram) vary across countries?
• How does the blocking of a service vary across countries and ASNs?
• How does the blocking of a service change over time?

When trying to answer questions like the above, we normally perform relevant data analysis (instead of inspecting measurements one by one).

The MAT incorporates our data analysis techniques, enabling you to answer such questions without any data analysis skills, and with the click of a button!

How to use the MAT?

Through the filters at the start of the page, select the parameters you care about in order to plot charts based on aggregate views of OONI data.

The MAT includes the following filters:

• Countries: Select a country through the drop-down menu (the "All Countries" option will show global coverage)
• Test Name: Select an OONI Probe test based on which you would like to get measurements (for example, select Web Connectivity to view the testing of websites)
• Domain: Type the domain for the website you would like to get measurements (e.g. twitter.com)
• Website categories: Select the website category for which you would like to get measurements (e.g. News Media for news media websites)
• ASN: Type the ASN of the network for which you would like to get measurements (e.g. AS30722 for Vodafone Italia)
• Date range: Select the date range of the measurements by adjusting the From and Until filters
• Columns: Select the values that you would like to appear on the horizontal axis of your chart
• Rows: Select the values that you would like to appear on the vertical axis of your chart

Depending on what you would like to explore, adjust the MAT filters accordingly and click Show Chart.

For example, if you would like to check the testing of BBC in all countries around the world:

• Type www.bbc.com under Domain
• Select Countries under the Rows
• Click Show Chart

This will plot numerous charts based on the OONI Probe testing of www.bbc.com worldwide.

Interpreting MAT charts

The MAT charts (and associated tables) include the following values:

• OK count: Successful measurements (i.e. NO sign of internet censorship)
• Confirmed count: Measurements from automatically confirmed blocked websites (e.g. a block page was served)
• Anomaly count: Measurements that provided signs of potential blocking (however, false positives can occur)
• Failure count: Failed experiments that should be discarded
• Measurement count: Total volume of OONI measurements (pertaining to the selected country, resource, etc.)

When trying to identify the blocking of a service (e.g. twitter.com), it's useful to check whether:

• Measurements are annotated as confirmed, automatically confirming the blocking of websites
• A large volume of measurements (in comparison to the overall measurement count) present anomalies (i.e. signs of potential censorship)

You can access the raw data by clicking on the bars of charts, and subsequently clicking on the relevant measurement links.

Website categories

OONI Probe users test a wide range of websites that fall under the following 30 standardized categories.