The data was gathered from three different sources and in three different formats as described below:

1. The WeRateDogs Twitter archive. This file was provided by Udacity and was downloaded manually by clicking the following link: twitter_archive_enhanced.csv and was saved as a csv file.

2. The tweet image predictions. This file was hosted on Udacity's servers and was downloaded programmatically using the Requests library and the following URL: image_predictions.tsv and was saved as a tsv file.

3. Retweet and Favourite count for each tweet. This file was queried from Twitter's API using the Tweepy library and the tweet IDs in the WeRateDogs Twitter archive. The data was stored in a text file file.

The three pieces of data were assessed visually and programmaticlly for quality and tidiness issues.

• Visual Assessment: Each piece of gathered data is displayed in the Jupyter Notebook for visual assessment purposes. Once displayed, data was additionally assessed in an MS Excel.
• Programmatic Assessment: Pandas' functions and methods was used to assess the data.

• Only original ratings (no retweets) that have images were kept. Though there are 5000+ tweets in the dataset, not all are dog ratings and some are retweets.
• The fact that the rating numerators are greater than the denominators does not need to be cleaned. This unique rating system is a big part of the popularity of WeRateDogs.

All issues documented while assessing data was cleaned in the "Cleaning Data" section in the wrangle_act.ipynb.

• Before cleaning, a copy of original data was made 8 During cleaning, the define-code-test framework was used and documented.
• The result was 2 high-quality and tidy master pandas DataFrames.

In the "Storing Data" section in the wrangle_act.ipynb notebook, the cleaned master DataFrame is stored as a csv file with name twitter_archive_master.csv. The addidional file was stored as twitter_archive_tweet_info.

Finally, visualizations were made on the cleaned data with matplotlib and seaborn after which the data was stored.

1. A written report called wrangle_report.html was created to briefly describe the wrangling efforts. This was framed as an internal document.

2. A second report called act_report.html was created to communicate all the insights through visualizations produced from the wrangled data. This was framed as an external document (blog post).

1. https://help.start.gg/en/articles/1987102-customizing-text-with-markdown#:~:text=Aligning%20Text,the%20text%20in%20div%20tags.

2. https://www.w3schools.com/PYTHON/matplotlib_pie_charts.asp

3. https://www.digitalocean.com/community/tutorials/pandas-merge-two-dataframe