Projects

CCD4DT

Visitor tracking for fairs and events

Track the position and path of trade visitors to analyze traction and gain insights into booth performance. To improve tracking accuracy, multiple sensor types can be combined (WiFi, RFID, camera). Together with a fellow UZH student, we developed a new system that correlates measurements of different sensors and predicts the actual position of visitors.

Expose REST API for ingestion of measurement data
Kalman Filtering to remove noise
Coordinate system standardization
Clustering and prediction
Persistence to InfluxDB and exposure over REST API

CCS4DT

GitHub logo Open in GitHub

Python / Pandas / InfluxDB / Flask

clinIQ

Real-Time Inventory Tracking for Hospitals

As a side project, I created the clinIQ project and founded dataloc GmbH. It’s mission is to help hospital staff find portable inventory (wheelchairs, beds, infusion pumps) within the hospital building to save time and increase efficiency.

Bluetooth Low Energy Beacons (tag inventory)
Crowd-sourced, distributed inventory tracking with mobile App (on staff’s smartphone)
Centralized processing of tracking data
Location prediction of inventory
Front-end to visualize inventory location

Bluetooth Low Energy / Python / MQTT / Laravel / React

NYC BnB Explorer

Explore the neighbourhoods of New York Airbnbs

Together with 3 fellow UZH students we developed a tool to explore and assess the neighbourhoods of Airbnb listings in NYC. We collected public datasets on NYC about crime, rodent, and noise complaints. The user can explore these datasets within the context of a particular Airbnb listing he/she is interested in.

Display Airbnb listings on a NYC map
Heatmap layer of crime incidents can be toggled on top of the map
Heatmap layer of rodent sightings can be toggled on top of the map
Heatmap layer of noise complaint incidents can be toggled on top of the map

BnB Explorer

GitHub logo Open in GitHub

Heroku logo Deployed to Heroku

Python / Flask / React

DeathNote

Smart contract to settle an inheritance

In the event of a death of a crypto currency holder, heirs can’t access the funds without the deceased’s private key. Together with 3 fellow UZH students, we developed a smart contract to address this problem (we were dead serious)

Smart contract on the Ethereum blockchain
Owner charges the smart contract and specifies:
List of heirs along with their shares
List of voters along with a policy (e.g. m/n votes required)
Voters can vote for the owner’s death. As soon as policy is met, owner is considered dead.
Without the owner’s objection, smart contract funds are released to heirs after 30 days

DeathNote

GitHub logo Open in GitHub

Heroku logo Deployed to Heroku

Blockchain / Ethereum / Solidity / React

BAZO

GDPR and its application and compliance issues with blockchains is gaining attention. During my Bachelor thesis I developed a solution that allows fine-grained update and erasure operations on transaction fields. A cryptographic primitive called “Chameleon Hashing” enables the efficient calculation of hash collisions. These collisions can be used to modify the content of transactions

Implementation of Chameleon Hash function library in Go
Integration of new hashing mechanism in existing BAZO blockchain and its protocol
Implementation of “update transactions” to modify on-chain data
Implementation of a web-based block explorer to visualize the results

Sequence diagram

Sequence of an UpdateTx

AccountTX (Before update)

UpdateTX

Account TX (After update)

GitHub logo Open in GitHub

Blockchain / Go / React

Santorini

We created a digital version of the game Santorini. The challenge on the front-end was to emulate the board game and provide the possibility to create buildings and move figures. On the back-end side, the challenge was to implement the game logic (e.g. rules and superpowers of the individual game characters).