Wikimedia Data Challenge
This Project was put out as data challenge by the Wikimedia Foundation for a data analyst position. I thought it would be a great opportunity to brush up on some data analysis tools in Pandas and decided to complete the challenge.
The goal of the project was to analyze data from event logging (EL) to track a variety of performance and usage metrics to help the company make decisions. Specifically, they were interested in:
- clickthrough rate: the proportion of search sessions where the user clicked on one of the results displayed
- zero results rate: the proportion of searches that yielded 0 results
The dataset comes from a tracking schema that the Wikimedia Foundation uses for assessing user satisfaction. Desktop users are randomly sampled to be anonymously tracked by this schema which uses a "I'm alive" pinging system that we can use to estimate how long our users stay on the pages they visit. The dataset contains just a little more than a week of EL data.
|uuid||string||Universally unique identifier (UUID) for backend event handling.|
|timestamp||integer||The date and time (UTC) of the event, formatted as YYYYMMDDhhmmss.|
|session_id||string||A unique ID identifying individual sessions.|
|group||string||A label ("a" or "b").|
|action||string||Identifies in which the event was created. See below.|
|checkin||integer||How many seconds the page has been open for.|
|page_id||string||A unique identifier for correlating page visits and check-ins.|
|n_results||integer||Number of hits returned to the user. Only shown for searchResultPage events.|
|result_position||integer||The position of the visited page's link on the search engine results page (SERP).|
The following are possible values for an event's action field:
- searchResultPage: when a new search is performed and the user is shown a SERP.
- visitPage: when the user clicks a link in the results.
- checkin: when the user has remained on the page for a pre-specified amount of time.
This user's search query returned 7 results, they clicked on the first result, and stayed on the page between 40 and 50 seconds. (The next check-in would have happened at 50s.)
The objectives of this project are to find answers to the following questions:
What is their daily overall clickthrough rate? How does it vary between the groups?
Which results do people tend to try first? How does it change day-to-day?
What is their daily overall zero results rate? How does it vary between the groups?
Let session length be approximately the time between the first event and the last event in a session. Choose a variable from the dataset and describe its relationship to session length. Visualize the relationship.
To install the requirements with pip (except for Python), type in the main directory:
pip install -r requirements.txt