The documentation of the camuflon project - WINNER
This project has been developed in just one day for the NOI Hackathon SFScon Edition (2021). With some fixes and feature implementation, could be an already made beta of a real product.
- Website: https://camuflon.vercel.app
- Api endpoint: https://camuflon-api.herokuapp.com
- Api documentation: https://camuflon.docs.apiary.io
- App: not available on playstore yet
"He is a large, peaceful ape with delicate movements. He has thick silvery fur, long, thin and silky, covering his sad black eyes. He has the ability to make himself invisible".
This is why we chose this name, because as a camuflon, our product makes employees invisible, while their anonymous tracks are still tangible and can be used to make statistics.
The purpose of the project was allowing companies to track where their employees pass most of their time, like in the office or the machines room. With this data many statistics can be calculated, for example to understand if rooms are not used and should be dedicated to other things. The data are aggregated or, in any case, anonymous, with the usage of a pseudo-identifier not correlated to any name.
Our solution is a software and companies could register to that service as a monthly paid service. It is already online, even if the site lacks several features, such as the signup (working in the API, though) and the rooms' deletion.
We used a sort of Agile method to develop it (of course with very tiny sprints) and this result could be thought as the result of the first sprint. With other two days, the website could be finished and some bugs would be likely fixed, so it would be already made for a beta. The API are already totally working and have a documentation site made with Apiary.
Our solution is composed of four parts:
- The APIs, developed with NodeJS and Typescript and hosted on Heroku
- The Database, with MongoDB and hosted on MongoAtlas
- The Website, developed with React and hosted with Vercel
- The Mobile App, developed with ReactNative and, hence, available both for IOS and Android
- The company signs up on the website, by putting company name (will be the username), email (will receive a confirmation email) and password (recovery is still to be implemented). The signup is available on APIs but not in the website
- The company logs in (for now, the credentials are: { companyName: "azienda", password: "HackathonBolzano2021!?" })
- The company does the initial settings (partially available): it can add beacons and rooms. To each room it can associate one or more beacons' ids.
- The company is assigned with a token that can see in the website (not visible yet, but present in APIs and DB). This token is given to employees, so that only employees can use the app to report statistics
- The employees download the mobile app and insert company name and token. They are provided by the APIs with an unique id, that will be saved in the local storage. Of course the APIs just check that the company exists and has that token, they have no way to identify the specific employee.
- If the employee changes phone, it can insert his old uid to preserve the old data and avoid adding a redundant new user.
- The mobile app gets the lists of available beacons' ids and starts detecting the nearest one. The app shows company name, user uid and the detected room. It shows also the exact beacon id for debug purposes.
- Every a certain amount of time (for this sprint it is fixed), the app sends to the APIs timestamp, location and user uid. This are the data used for the statistics.
- If he wants, the user can pause the data sending. If the beacon detection is not working, it can receive notifications such as "enable your bluetooth" (not implemented yet).
- The company can login and see statistics. For now they are only average weekly room usage and how many minutes are spend in each room on a certain weekday. These are just the top of the iceberg and many other statistics could be implemented with more time. For instance, machine learning could be used to find patterns such as "Users that use the kitchen in the morning, spend far less time in the machines room than users that don't use the kitchen". Machine learning and additional statistics could be implemented in a second or third Agile sprint.
- The employee can use the site as well, he has no credentials but he can put company name, company token and uid to see just his own data, by remaining anonymous.
In this first sprint, free services have been used:
- MongoDB Atlas for the database
- Heroku for the api hosting
- Vercel for the website hosting
In order to have a better product, in other potential sprints, Docker images could be developed for both API, website and database. The API Docker Image has already been automatically developed with a Github Action and is available here on dockerhub.
Kubernates could be used to serve them.