This aims to be a dashboard that predicts trends in fashion, based on previous garments that have been worn and photographed on runways, as well as similar pieces seen on other lower brands (like fast fashion brands) and similar items seen on social media.

The problem with the fashion industry today is the overproduction and waste of materials through fast fashion, failed sample garments, etc. causing environmental damage in our society. This forecasting tool could help designers better predict consumer preferences and demand through deliberate trend adoption using data-driven predictions and timelines. This will enable more precise production planning and help slow down the cycle. For this prototype a data-set would derive from a select range of sources (i.e. fashion magazines, fashion week achieves) and use image and text data-types to identify "fashion elements" (e.g. patterns: leopard print).

The dashboard will have a trending analytics panel showing graphs, of popular fashion elements from previous years and the current year, pie charts of popular fashion elements in particular sources (e.g. magazines, fashion brand websites) etc. There will also be a graph indicating the forecasting trends of the season. Time series forecasting may also be required here; timestamps of when garments from collected data were released. It will open up on the homepage showing top 5 emerging trends, with a layout of images and hashtags etc.

Forecasting will require robust AI models and machine learning: computer vision, data reprocessing, Natural Language Processing, APIs, visualisation tools (like D3.js) for the interactive dashboard and a database system (most likely using SQL).

It would be targeted to fashion brands, independent designers, marketing teams etc.