NovelGens Data Analysis

Data source

The data consists of website data provided by Indent Labs' NovelGens, which generates fictional creatures using AI.

Business problem

Indent Labs used artificial intelligence to create 365 creatures in 2022 — one new, unique creature every day. Each creature was assigned a commonality, biome, tag, and paragraph of content. Users then had the ability to claim the creature, like the creature, create artwork of the creature, and/or write a story about the creature.

The website collected anonymous feedback from users about each creature, asking: Is this creature memorable? Scary? Sad? Powerful? Unique? Comical? Comforting? Dangerous?

After producing creatures and collecting data for a year, Indent Labs seeks to better understand the trends between the qualitative data and the user-submitted feedback. They want to understand:

• Which creatures claim superlatives in each of the feedback categories? What factors contribute to this?
• What trends exist between creatures that share types, tags, and/or environments? What is the most common feedback associated with each of these data points?
• What factors influence users to interact with each creature, either by claiming it, liking it, or submitting new content about it?

Data cleaning

I received 20 .csv files from Indent Labs. I cleaned them using Google Sheets, eliminating unnecessary data, filling empty data, and reorganizing their structure before importing 16 tables to PostgreSQL.

1. content_favorites: The "likes" given to each creature, consisting of 5 columns and 408 rows.
2. content_feedback: User-submitted feedback about each creature, consisting of 18 columns and 4,652 rows.
3. creature_environment_pairings: The environments selected for each creature, consisting of 5 columns and 655 rows.
4. creature_story_pairings: The user-submitted stories and their respective creature(s), consisting of 5 columns and 20 rows.
5. creature_tag_pairings: The tags selected for each creature, consisting of 3 columns and 527 rows.
6. creatures: Published and drafted creatures, consisting of 8 columns and 369 rows.
7. environment_story_pairings: The environments selected for user-submitted stories, consisting of 5 columns and 2 rows.
8. environment_submitted_image_pairings: The environments selected for user-submitted images, consisting of 5 columns and 38 rows.
9. environments: All possible environments for creatures, consisting of 5 columns and 18 rows.
10. ownership_contracts: Creatures that have been claimed by users, consisting of 6 columns and 3,050 rows.
11. profiles: Profile pages for website users, consisting of 7 columns and 154 rows.
12. submitted_content_kudos: Kudos given by users to submitted stories and images by other users, consisting of 7 columns and 135 rows.
13. submitted_images: Images submitted by users about the creatures, consisting of 9 columns and 53 rows.
14. submitted_stories: Stories submitted by users about the creatures, consisting of 7 columns and 20 rows.
15. tags: All possible tags for creatures, consisting of 5 columns and 98 rows.
16. users: Encyrpted data on website users, consisting of 9 columns and 154 rows.

Contact info

Please reach out to me if you have any questions or feedback.