0
$\begingroup$

I have a data concerning the number of posts about certain hashtag on Instagram by influencers and not influencers per hour. The time period is one week, the dataset looks like this:

Hours 2 PM 3 PM 4 PM 5 PM 6 PM 7 PM
Influencers 8 15 10 20 13 25
Not influencers 38 58 30 28 49 75

I want to perform some statistical test to figure out is there a different in the patterns of posting between two groups. For instance, can we figure out which group is this just looking at the pattern of the distribution. But I am confused which kind of test I need to perform. What kind of test can I choose taking into account that data is not distributed normally and probably one group can influence another group?

Improve this question
$\endgroup$
2
  • $\begingroup$ I think you should first test for autocorrelation in both categories, because observed data measured over time do not necessarily need to be treated as time series. That is, next values may be unrelated to the previous ones. If you can treat each category as iid samples you can do distribution fitting. Not posting this as an answer because it is not thorough enough. $\endgroup$
    – Ghostpunk
    Commented May 14, 2023 at 15:07
  • $\begingroup$ To find differences in patterns you don't need a test: you need to explore the data. Start by graphing them in a way appropriate for the kind of patterns you are looking for. If ultimately you expect to formulate definite hypotheses and test them, split your data into a training group for exploration and a complementary testing group for evaluating the hypotheses you develop. We can't tell you much about what tests you might consider because they will depend on the hypotheses you make, what you learn about the data, and your reasons for the analysis. $\endgroup$
    – whuber ♦
    Commented May 14, 2023 at 15:55

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.