background
Indigo Jamis is a demo web application providing browsing, exploration and analytic capabilities for predicting and investigating viral videos.
The primary goal of the product is to demonstrate cutting edge machine learning and analytic capabilities to a potential client.
process / requirements
User task requirements
- Triage flow
- Collection flow
- Drill-down flow
- Collection flow
- Drill-down flow
Content requirements
- Interactive video thumbnail
- Predicted virality
- Predicted view count
- Video metadata (poster, date posted, comments and tags)
- Interactive video thumbnail
- Predicted virality
- Predicted view count
- Video metadata (poster, date posted, comments and tags)
In the first ideation phase I drew on features of existing tools and applications in an effort to provide a familiar browsing experience to the user. Annotated whiteboard sketches are shown below.
The first mockups for Indigo Jamis featured a custom search query box and visual representations for the underlying analytics. I found this version uncompelling, with not enough emphasis on the analytic aspects. The visualizations were also difficult to parse in a quick glance triage use-case.
solution
Indigo Jamis mockups include grid, list, and single item views. Multiple views provide a flexible browsing experience for the user.
default grid view
bookmarked search grid view
bookmarked search list view
bookmarked search single item view
viral item detail view
non viral item detail view
Each video card exposes the predicted virality, views, and content analytics. In this version the analytics are easily available at a glance; a flame icon is either colored or not to indicate predicted virality. Hovering over the video thumbnail cycles through important video frames. Other metadata such as poster, upload date and video length are also included in the summary card view, with full metadata available in single item view.
I also chose to include custom query and bookmark functionality as a stretch goal. Though we primarily want to highlight the underlying analytics, I also want the prototype to feel like a complete application that the client can see fitting into their analyst workflow.
presentation / demo video
The prototype was presented to the client by way of recorded prototype video and live demo. The screens in the video below are taken from the MVP prototype of Indigo Jamis.