Openhardware Social Network Analysis in Mastodon

The traffic analyzed was extracted from the Mastodon instance infosec.exchange, an instance about info/cyber security-minded people.

First of all, all the accounts were extracted from the 4505 list using the Mastodon API with the following bash command:

curl -s -k -H 'Authorization: Bearer ' -X GET 'https://infosec.exchange/api/v1/lists/4505/accounts?limit=0' | jq '.[] | .acct' | sed 's/\"//g' > accounts_openhardware.txt

Then, the following command allow us to extract any post from the above accounts:

cat accounts_openhardware.txt | while IFS= read -r line; do sh get_statuses_02.sh $line | jq '.[] | [.content,.reblog.account.acct, .reblog.content]'; done > content_openhardware.txt

>> get_statuses_02.sh
acc=$(echo $1 | cut -d '@' -f1)
server=$(echo $1 | cut -d '@' -f2)
id=$(curl  -s -k -H 'Authorization: Bearer ' -X GET "https://"$server"/api/v1/accounts/lookup?acct="$acc | jq .id | sed 's/\"//g')
curl -s -k -H 'Authorization: Bearer ' -X GET "https://"$server"/api/v1/accounts/"$id"/statuses/?limit=100000"

Once we have all the data inside the content_openhardware.txt file, it is necessary to parse its content.
NOTE: jq command returns data in a poor format, it is recommended to edit the file and add commas between each extracted element.

The capture below compares the extracted terminal output (left) with the cleaned JSON dataset (right). Each dataset entry contains cleaned text and extracted entities; the commands above remain available as copyable text.

Enlarge

This final dataset contains an element for each post with the text cleaned (parsing the HTML format and extracting any email, url, mention, hashtag and emoji). Also, a list of matches from different types (EMAIL, URL, MENTION, HASHTAG and EMOJI) with each tag found attached with its position in the original text.

Social Media Insights: Trending hashtags, mentions and domains

We’ve created some wordclouds and pie charts to illustrate the most prominent elements between hashtags, mentions, and domains.

The Hashtags Word Cloud vividly displays the range of topics buzzing in conversations. Dominating this landscape are #opensource, #linux, and #orconf, as seen in the pie chart, indicating a strong focus on technology and innovation among users.

The Mentions Word Cloud reveals key influencers and active participants in the digital dialogue. Among them, @ngizero and @oshpark emerge as the most mentioned profiles, suggesting their significant impact on discussions, as corroborated by the pie chart.

Lastly, the Domains Word Cloud underscores the most referenced websites, highlighting the sources shaping public opinion and information flow. Notably, github.com and youtube.com lead the chart, reflecting their central role in content sharing and community engagement. However, ignoring these two pages we can highlight fosslife.org or riscv.org as the most shared pages in the openhardware community.