close questionmark user search earth triangle retweet users star home mention pencil info lightbulb-o icon_fullscreen images refresh list grid mail in eye export twitter

Creating the Social Bearing website

05 January, 2015 by Tom

Social Bearing is a Twitter based statistics website which I launched a few days ago. I started developing the site in May 2014 and it combines my interests of statistics, code and enthusiasm for the Twitter API.

Social Bearing Screenshots

Screenshots showing some of the features on social bearing including open search, follower view and user timeline search

The website is currently in BETA, so there are plenty of features I still want to develop over the coming months.

Evolution of an idea

Any visitor to my webdevdoor.com blog may have noticed a few articles I’ve writted on the public Twitter API. At the end of 2012, I wrote about creating a custom Twitter feed that people could put on their websites. When Twitter brought in the authentication requirements for all streams, I wrote about basic authentication and have since written about caching Twitter streams, Twitter feed animation and other aspects of the API.

At the beginning of 2014, I launched a fairly simple project for a Twitter Job Search tool. Although it links externally through to the official Twitter Search, it gave me further insights into how Twitter search queries operate and how geo-located tweets can be found.

Social bearing builds on much of these ideas. When I started the project, it was initially only going to be an open search and sentiment tool for public tweets. As I started building the site and getting into the nuts and bolts of the twitter streams, I found further features I wanted to add including people search, home feed display and follower/friend analysis.

Twitter Tool Research

Looking at the competition and other Twitter tools was a big part in helping me understand the direction I wanted to take Social Bearing. Some of the more popular tools I evaluated include TweetReach, Topsy and Twitonomy.

The more research I did, the more I felt my approach with Social Bearing was different. I wanted to create something unique and believe I have achieved that by how tweet and user entities are displayed, summarised and filtered.

Most Twitter tools for example won’t allow you to sort tweets and have limiting filtering. Social Bearing can sort tweets by metrics including engagements, reach, followers and sentiment. Tweets can also be filtered by content type (pictures, links, mentions etc), language and top words, contributors and influencers.

Social Bearing also displays every tweet (excluding possibly sensitive tweets) that matches the search query and gives the user the option to filter less relevant tweets as they wish. Many Twitter tools seem to automatically filter the display of tweets to greater degrees before displaying to the user.

Current Features

Currently Social Bearing includes the following main areas:

  • Search and analytics for public tweets
  • User timeline search
  • People search
  • Followers and friend stream
  • Personal statistics for the signed in Twitter user’s home feed, mentions, favorites and retweet stream

Each of these main areas will display a summary of the data and have the ability to filter and sort users and tweets by metrics such as engagements, reach, influence and language.

Potential uses could include:

  • Finding interesting tweets on any subject
  • Hashtag, brand and trend monitoring
  • Reputation management
  • Providing insights into anyone’s timeline
  • Providing insights for your home feed including retweets and mentions
  • Finding interesting people to follow
  • Filtering out friends to unfollow

Some potential features

I have a lot of new features and ideas I would like to develop for Social Bearing in the coming weeks and months. Some of these include:

  • Finding geolocated tweets. I think it would be particularly useful for people to be able to search for any activity on Twitter based on the actual location people are tweeting from.
  • Map view for geolocated Tweets. An interactive map of Tweets and tweet types would be a great way to visualize at a glance where people are tweeting from.
  • Individual tweet statistics
  • Ability to save or ‘pin’ tweets
  • Better language selection

It would be great to hear your feedback or any suggestions for the site. 🙂



3 Comments

  • Sebastian says:

    Hi guys,

    it’s a great tool in my opinion! I use it for a short Twitter analyses for my master thesis.

    Just one question: My search resulted in 233 tweets within a timeframe of 9 days. But it’s written that only the last 7 days are shown. So, my question is: Are the 233 tweets within the timefrome of 9 days or in the max. period of 7 days? Moreover, the counted tweets at “Tweets Over Time” sums up only to 225 tweets.

    Hope to receive an answer 🙂

    Best regards,
    Sebastian

    • Tom says:

      Hi Sebastian,

      Thanks for your feedback and pointing out these discrepancies.

      Regarding the timeframe for the appearance of tweets, it is true that the Twitter API for search and other sources mention a 7-day limit for the search index, which is why you’ll also see that mentioned on Social Bearing. However, I have noticed that sometimes it does display a timeframe of 8 days, and even on occasion 9 days.

      When this happens and you open the last tweet in the results direct on Twitter, it should correspond correctly to the timeframe displayed. e.g. placing a search today (09/01/2016) and viewing a tweet direct on Twitter from ‘8 days ago’ that’s returned in the Social Bearing search should have a date of 01/01/2016.

      Twitter therefore must sometimes be returning tweets older than 7 days. I suspect the reason for this is to take into account the 24 hour difference in location people are viewing tweets from, but I haven’t been able to find a definitive source to back this up.

      Regarding the counted tweets over time graph, this is a bug with the graph plotting that we will try and fix. The total tweets displayed at the top will give you the true total.

      Hope that helps and best of luck with your thesis!

      Tom

  • Paul Hinrichsen says:

    Hi Tom.

    Congratulations on a FINE piece of work. I have never seen anything else ]quite like it. The stats shown is truly amazing !

    Paul

  • Leave a Reply

    Your email address will not be published. Required fields are marked *

    css.php