Smartrobe — Reduce your closet and finally have something to wear.
This project was carried out as part of the TechLabs “Digital Shaper Program” in Dortmund (summer term 2023).
In a nutshell
Smartrobe aims to cut mindless consumption of clothing by analyzing your closet. Join us and finally have a smart wardrobe with only the things that you really love and need! Create a digital twin of your current closet by adding all of your clothes with corresponding details in the app. Each time you wear a piece, mark it as worn. Over time you’ll gain insight into your favorite pieces and outfits, cost-per-use and what probably to get rid of.
Background of the project/ Problem
Fashion is one of the biggest markets in the world. “Up to 100 billion garments are produced by the fashion industry every year, each year, as much as 92 million tons of clothing ends up in landfills” (source: https://theroundup.org/textile-waste-statistics/). Although this leads to an unbearable amount of waste, resource usage and toxic emissions, clothing companies still create new collections of clothes in huge quantities — with the latest trend of ultra fast fashion — convincing people that they need them. Meanwhile, our own closets are already packed with “must haves” from previous seasons, many of them rarely worn… That makes us part of the problem and responsible to act.
Just buying less clothes sounds much easier than it is, especially, if we don’t know what we exactly like. The main purpose of this app is to help people realize what they really wear and in consequence buy less unnecessary clothes.
Smartrobe aims to cut mindless consumption of clothing by analyzing your closet. Join us and finally have a smart wardrobe with only the things that you really love and need! Buy less (often useless and incompatible) new clothes, safe time, space and money and make your wardrobe a Smartrobe.
Solution
Having a big collection of clothes makes it very difficult to track what we are actually wearing. Oftentimes we are not aware which clothes we actually wear on the regular. This can make clearing out the closet an overwhelming and unmanageable task. Having it in an application, which shows with a few clicks statistical numbers about how many times a specific item was worn, makes it much easier to control our wardrobe and cater it to everyone’s individual needs. Part of this solution is also to make a kind of a “hitlist” of the clothes most worn to make decision making easier for the user. An additional part of our solution approach is to show the user the cost per use of a specific item. This way the user has concrete figures on how much use an item actually has for them. A cheap item that is never used might actually be just overconsumption and not needed while one expensive jacket might in the end give the user much more of their money worth.
Methodology
We started by brainstorming ideas for our problem. We first collected a lot of those ideas that later had to be evaluated on how implementable they were in the time we had and which ones were the most essential. The next big question was if we wanted to separate the frontend and the backend or keep everything in the frontend. There were different reasons for and against it, but since all of us worked on different tracks the last month and learned to use different programs we decided that it would be best to separate back- and frontend. As the members of the team have very different backgrounds and are scattered across the Ruhrgebiet, we quickly adopted a remote and agile workflow to progress in the project Communication took place in slack — with live meetings and asynchronous chatting. Persistent information, tasks and a small “wiki” were created in Notion.
To ease setup and get people from different expertises start the project, we decided on containerizing development right from the beginning.
The app itself is designed upon a RestAPI approach with a fastAPI backend that handles data and statistics (the actual data being stored in a SQLite Database) and a React Frontend for UI and Interaction.
Summary:
- Organization: agile workflows in Notion, Slack;
- Infrastructure: git, docker;
- Frontend: Languages: HTML, CSS, TypeScript; Tools: VS Code, React Library;
- Backend: FastAPI, Pandas, SQLite (language:Python).
- UX Design and User Research: Figma, Quantitative methods (Survey)
The final project of Smartrobe:
https://www.figma.com/file/cvU5sDrYhWuBrmSAWrcajf/Smartrobe-Project?type=design&node-id=220 3%3A1148&mode=design&t=cNnqwnUa2mK4SLuf-1
The project was changing very dynamically, this is the reason why there are several differences between project and final product design.
Results
By now the app is in its MVP state and consists of a web application with its core features which are creating a digital twin of a closet by adding all clothing pieces and basic wearing statistics. The pieces contain a description and information about size, color, brand and price.
Besides the information about the wardrobe, a core part of the app is to gain insights into one’s own using and wearing behavior. As a first metric “cost per use” is implemented, to evaluate if the initial investment into that exact piece of clothing was already worth it.
Next steps in the project development
Due to time limitations we couldn’t implement all of the ideas that we had. There are many features which could be added to the project to make it more useful and attractive.
Below we present some of the ideas for future development:
- explicit overview page for each piece -the possibility of adding pictures of clothes to the closet;
- categorizing clothes not only according to type/purpose but also according to the season;
- the function of adding different categories by the user, according to individual needs;
- the possibility of adding clothes that were owned by the user for a longer time, with frequency of wearing based on the users memory;
- filters for faster and more effective searching of specific pieces (according to color, brand etc..);
- statistics section improvement(e.g., deeper insights about what percentage of the closet and pieces in particular are regularly worn, preferred colors);
- recommendations of what can be done with unwanted clothes (e.g. information about where to donate/sell clothes or about clothing exchanges in the area)
- data about sources statistically used for the production of specific pieces of clothing (such as water and energy)
Explicit long term goals of the application are:
- A mobile native version to be really useful in everyday life, especially in adding pictures;
- AI assisted entry of pieces and recommendation algorithms.
For visualization purposes, below we present a few sheets from a project made by our UX designer, Ayse Nur Eren.
GitHub repository (or similar):
Team members:
Web Development:
- Paulina Daciuk [paulinadaciuk@gmail.com, www.linkedin.com/in/paulina-daciuk-494731134, frontend]
Machine Learning:
- Jannik Nahrgang [jannik.nahrgang@googlemail.com, https://www.linkedin.com/in/janniknahrgang/, architecture and backend]
- Ayman Soultana [ayman.soultana@ruhr-uni-bochum.de, Ayman Soultana | LinkedIn, architecture, backend, frontend)
UX Design:
- Ayse Nur Eren [aysenurerenn9@gmail.com, https://www.linkedin.com/in/aysenurkocak/, user research, ux/ui design)
Data Science:
- Erik Ruslanov [erik.ruslanov@gmail.com, data frames, backend]
- Marlena Stojakovic [marlenast11@gmail.com, backend]
Team mentors
- Tom Stein, connection of frontend and backend, docker support
