Episode 2 of #satyamevjayate sees over 50% drop in mentions on Twitter compared to debut episode.

By vishesh on May 14, 2012

We received great response to our analysis of Satyamev Jayate mentions last week and you must be curious to see what happened yesterday.

  1. Let me give you the top-line – conversations on Twitter saw a 52% drop over last week and we noted a total of 12565 mentions of #satyamevjayate (only this specific hashtag)
  2. In terms of timeline, the pattern was very similar to the 1st episode, though volumes were much lower
  3. Volumes matched upto 11 am, but couldn’t build-up to last week levels thereafter
  4. Note the significant drop in reach and impressions, showing a drop amongst the highly connected

Here is this week’s report:

Twitter report star plus satyamev jayate_13 may, 2012

View more presentations from Blogworks
Posted in Social Media | Comments (1)

Analysis of today’s Twitter conversations around Aamir Khan’s #SatyamevJayate

By vishesh on May 6, 2012

The build-up for Aamir Khan’s Satyamev Jayate has been unprecedented and a phenomenal 16700 posts were generated today (upto 15oo hours IST) as the show opened today. We decided to take a peek into Twitter conversations today. Hope you find this useful.

Do share your thoughts on what you felt about the conversations today.

Twitter report star plus satyamev jayate_6 may, 2012

View more presentations from Blogworks.
Posted in Social Media | Comments (3)

IndiaSocial Summit 2012 Twitter Analysis #indiasocial12

By vishesh on April 14, 2012

IndiaSocial Summit 2012 evoked a fair amount of buzz and conversations online generating over 6000 messages between Jan-April 2012 (of which over 3800 were shared between 2-4 April 2012 – the event duration).

Here is a part of the report we prepared for our internal assessment.

IndiaSocial summit 2012 twitter analysis by team blogworks

View more presentations from Blogworks

Team Impact Measurement also shared a report, which makes for a great resource too (Thank you) – given that we have used different tools, results show some variation.

Posted in Blogworks, Social Media | Comments (0)

Viralheat’s Twitter sentiment analysis plug-in for Chrome: my feedback

By vishesh on October 16, 2011

Automated sentiment analysis is one of the most criticized features within Social media analytics / monitoring family.

From my experience I have an idea that out of all comments generated by any tool, be it Radian6, Alterian SM2, or any other tool, there is an accuracy gap. I believe that the inaccuracy of measuring sentiment by any tool is about 70%. Most of these gaps occur when the sentiment is marked neutral.  Another reason of gap is these tools are not able to identify local slangs, words or phrases translated or written in English, sarcasm and mood of author.

When I recently read the news that Viralheat, a social media monitoring tool, has released free Chrome plug-in that adds sentiment analysis to Twitter timeline, I decided to give it a shot.

About Viralheat Sentiment Analysis Plug-in for Twitter:

Viralheat’s this new plugin can be found easily on Chrome web store and it takes couple of minutes to install. Once the plug-in is installed, and you open your Twitter page on Chrome, the plug-in adds a bar on top of Twitter page and displays aggregated sentiment for displayed tweets on that page. When user searches for any brand i.e. ‘iPhone’, it will provide sentiment for all relevant tweets.

What’s New and Good?

My first thought was that the tool is pretty handy if I am on my Twitter page. It lets me see the sentiment of the post and get overall idea of what’s happening around the brand, product or service I search for. One other feature I found interesting was that it lets you change the sentiment on Twitter page itself.

This tool can prove handy to see the sentiment around a live event or a hashtag (#).

What’s Bad?

I thought to run a small test on the accuracy of the tool. I took a sample of 50 tweets and compared the sentiment provided by the tool with manually analysed sentiments for same tweets.

I noticed, the tool analysed multiple posts on same topic with different sentiments. For example, a news article which was shared by many users, “Samsung Galaxy SII sales hit 10 Million, excluding United States” was analysed incorrectly 85% times. The overall accuracy was observed to 30%.

My Rating

3/10

Posted in Social Media | Comments (0)