Twitter Investigation

my hypothesis:
I predict women will be more emotional over twitter, i will measure this using the word count and mixed modality

Methodology:
- women show more emotion
- longer tweets
- 10 random male and 10 random female tweets
- because its a small pool I had to randomise it to limit the bias in the data

Data:

feature                             girls                          boys                                 difference

non standard grammar      2                                7                                        5

emojis/ tech                      12                               5                                        7

words                                76                              35                                      41

percentage of
emojis used in                  16%                           14%                                    -
overall word count

Analysis:

my hypothesis is supported by a limited pool of data. For example 16% of the word count in female tweets are of mixed modalitycompared to 14% in male tweets only 2 percentage point difference this is limited however because I had limited tweets. The word count also may have effected my data as female tweets had 41 more words 85% higher than male tweets making the data difficult to judge.


Conclusion:

To conclude my hypothesis in the begging was irrelevant so i had to adapty it, when i had adapted it my data began to give me some results that slightly supported my hypothesis however because it is such a small pool of data and some of the data were retweets it may have been bias and inaccurate.

for the future:

Research first!
before deciding hypothesis and data
make sure your data cllection is precise and you limit flaws.