Brand and Sentiment Analysis of Twitter Influencer: Dr Dipo

This project is on brand analysis of a Twitter influencer Dr Dipo

Brand analysis is an exploration that gives insight into how a brand is perceived by its audience. A brand could be a person, a company or a product

This particular analysis was done by extracting user’s tweets from Twitter (and retweets) for three weeks and performing an exploratory data analysis and a sentiment analysis of the tweets to evaluate the perception of the audience (in this case, Twitter users) about the brand (which in this case is Dr Dipo.

Sentiment analysis (or opinion mining) is a natural language processing technique used to determine whether data is positive, negative or neutral. Sentiment analysis is often performed on textual data to help businesses monitor brand and product sentiment in customer feedback, and understand customer needs.

Twitter sentiment analysis allows you to keep track of what’s being said about your product or service on social media and can help you detect angry customers or negative mentions before they turn into a major crisis. This can be done by picking the frequency of key words after removing stop-words or by evaluating tweet likes to

The Sentiment analyzers used for this project is a Vader Sentiment Analysis, you can read about it here.

Tweets from Monday, 2021-03-15 to Sunday, 2021-04–04 were extracted and analyzed using these search words: @ogbenidipo, Dr Dipo, Ogbeni Dipo. These search words and tweepy only gets quite a really good number of tweets as some users might have used these search words in their tweet.

So fortunately and fortunately Dr Dipo trended on 2021–03–24 because of the giveaway saga that happened that day and that makes this analysis even more interesting. Also, bear this in mind as you read through.

Results of the Analysis

A total of 1421 unique tweets were extracted (duplicates exclusive)using the direct Twitter API and Tweepy. Over the course of three weeks (3 to balance out even numbers), these are the insights I got

1. The top 10 usernames that tweeted most often during this period were:

· PurpleIntestine 14

· fuski_ 12

· Stan_Europe 11

· tikobadman 6

· IconicNino 5

These 5 people tweeted the most, why and who are they? perhaps Dr Dipo might know.

2. Nigerian location where most tweets came from, state or capital

Locations where the tweets came from

From the chart, it is observed that people from Northern Nigeria didn’t tweet much; I really thought Lagosians would dominate the tweets but it’s seen as otherwise ( someone said Ibadan people like too giveaway, Lol, pardon me please).

I think Dr Dipo should invest more in Ibadan people, then, the people in PH, I do love those people.

3. Top 5 influencers and their follower_count that tweeted about Dr Dipo or engaged other people’s post.

· MisterRedefined 88876

· PoojaMedia 74969

· BhadmusAkeem 74441

· TheOladeile 71092

· wakawaka_doctor 67371

These are the people with influence that Dr Dipo should get a little bit interested in, I don’t know if he knows them already.

4. Account age distribution of all the tweeters

Account age distribution

Most of the tweeters created their account in 2020 and 2019, this explains why a lot of the people that do engage with Dr Dipo’s post are youths.

4. Time of day the people tweeted the most

Hourly Tweeting frequency for three weeks

So these people tweet the most at 9 am, next is 12 pm and finally, by 8 pm, a brand could use this tweeting time to target its audience.

5. Day of the week the people tweeted the most

Daily Tweet frequency

There seems to be no pattern here, the tweet frequency is determined by how many tweets Dr Dipo drops for that day.

But on a general note Saturday seems quite positive as the Thursday in week 2 is actually the remnants of the giveaway saga which I removed already; then Friday follows showing a consistent pattern.

6. The sentiment analysis of the tweets

Creating a histogram from the compound of the sentiment scores gives this:

It seems like the people were a little bit unhappy or neutral after all, and it shakes off after a while.

For the techy guy, you can check the codebase here

Thank you for reading, I’ll be happy to connect on Twitter or engage in any new project.