New Launch: OONI Measurement Aggregation Toolkit (MAT)
Today the Open Observatory of Network Interference (OONI) team is
thrilled to announce the public launch of the OONI Measurement Aggregation Toolkit (MAT)!
The MAT is a tool that enables
you to create your own custom charts based on aggregate views of
real-time OONI data collected from around the world.
Use the MAT to track internet censorship worldwide based on
real-time OONI data!
About the MAT
Are you a researcher, journalist, or human rights defender interested in
investigating internet censorship around the world? OONI’s new
Measurement Aggregation Toolkit (MAT) was built for you!
The MAT enables you to track internet censorship around the world and
create your own custom charts based on real-time OONI network
What is OONI data?
Since 2012, OONI has built a free and open source app (called OONI Probe) that you can install and run to
measure various forms of internet censorship. Through the OONI Probe app, you can run tests on your network to
measure the blocking of websites, the blocking of
instant messaging apps
(WhatsApp, Facebook Messenger,
Signal), and to measure the
reachability of censorship circumvention tools (such as
Tor), among many other
tests. By default, all OONI Probe test
results are automatically sent to OONI servers, processed, and
published as open data in real-time.
In short, OONI data is OONI Probe test results. And OONI Probe test
results contain network measurements on the testing of websites and apps
around the world, often containing evidence of internet censorship.
Since 2012, OONI has openly published more than 808 million network measurements collected from 24 thousand
networks in 240 countries and territories. Every minute, as OONI Probe
users around the world run more tests, new measurements are
automatically published in real-time. The OONI database is therefore
one of the largest open datasets on internet censorship to date.
Why use the MAT?
To enable the public to explore OONI data and to investigate internet
censorship around the world, we created OONI Explorer, which is a web platform that
enables users to explore OONI measurements through a search tool.
However, OONI Explorer lists individual measurements, requiring users to
inspect each measurement one by one. This can make it difficult to
observe trends and patterns, as identifying and confirming the
blocking of services often requires looking at many measurements at
once (“in aggregate”).
We have therefore provided data analysis support to our partners and
broader community over the years, analyzing OONI data in order to answer
key questions such as 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. LGBTQI websites) are blocked in each
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
How does the blocking of a service change over time?
The MAT incorporates our data
analysis techniques, enabling you to answer such questions without any
data analysis skills, simply with the click of a button! You can also
generate your own custom charts based on aggregate views of real-time
OONI data, which you can then share on social media and use as part of
your research and/or advocacy.
Using the MAT
The MAT provides filters that
enable you to select the parameters you care about in order to plot
charts based on aggregate views of OONI data.
Specifically, 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.
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
Date range: Select the date range of the measurements by
X axis: Select the values that you would like to appear
on the horizontal axis of your chart
Y axis: 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 to generate a chart.
For example, if you would like to check the testing of BBC in all
countries around the world:
This will plot numerous charts based on the OONI Probe testing of
www.bbc.com worldwide, as illustrated below.
Interpreting MAT charts
The MAT charts (and associated
tables) include the following values:
OK count: Successful measurements (i.e. no sign of internet
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
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), 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
For example, if you would like to check the potential blocking of
Instagram in Russia through the MAT:
This will generate the following MAT chart.
From the above chart, we can see that
previously accessible in Russia, and only started to present signs of
blocking on 13th March 2022. From 14th March 2022 onwards, we observe a
large volume of anomalies which has persisted ever since, providing a
strong signal of blocking. We are also able to see that the blocking of
Instagram is confirmed on some networks (where measurements show that
the blocking of
www.instagram.com was automatically confirmed).
Moreover, the fact that some measurements over the last month show that
www.instagram.com was accessible suggests that the block is not
implemented on all networks in Russia. To explore how the blocking of
www.instagram.com varies across networks in the country, select
ASN under the
Y axis of the MAT.
Please note that the presence of anomalies is not always indicative of
blocking, as false positives can
occur. A large volume of anomalies can provide a signal of potential
blocking, but it’s useful to corroborate such data with other relevant
resources (when possible). For example, when looking at measurements
pertaining to the reachability testing of circumvention tools like
Psiphon, it’s worth looking at
Tor Metrics and the Psiphon Data Engine (both of which share usage stats) as well.
Quite similarly, OONI data on the testing of
www.youtube.com can be
cross-referenced with YouTube traffic data from the Google Transparency Report.
OONI data on the testing of websites can also generally be compared with
data collected from Censored Planet.
You can access raw OONI data by clicking on the bars of
MAT charts, and subsequently
clicking on the relevant measurement links.
We hope the internet freedom community finds the
MAT useful for investigating
and responding to internet censorship events around the world.
If you have any questions, feedback, or feature requests, we’d love to
hear from you! You can open a ticket or write us an email.
We thank all OONI Probe users worldwide
who have contributed (and continue to contribute) measurements,
supporting the MAT. We also
thank our community members for their invaluable feedback.