The online supermarket Picnic uses the volume of groceries to pack them into totes and bags. While this generally works well, there are cases in which an article's volume fits into a designated space, but its dimensions do not. To prevent this so-called overflow, Picnic uses a maximum fill rate of 85%, which means that no tote can be filled to more than 85% of its volume. Unfortunately, the use of a maximum fill rate does not guarantee that no overflow occurs at all and workers are sometimes still instructed to perform infeasible packings. To solve this, this thesis proposes a Three-dimensional Bin Packing algorithm (3D-BAGS) which uses the width, height and length of an article to determine whether it can be placed inside a bag. A Biased Random-Key Genetic Algorithm is used to converge to a good packing solution. After which, the same algorithm divides the bags into so-called totes, which are shipped to the customer. 3D-BAGS expands upon the state-of-the-art Three-dimensional Bin Packing algorithms by including bag stretching plus the diagonal rotation and squeezing of articles. Computation time of the current state-of-the-art is reduced by including memoization, priority based multithreading and a different stopping criterion. This thesis shows that the use of 3D-BAGS leads to more accurate packings, while simultaneously lowering the amount of totes needed to ship ambient and chilled groceries. However, 3D-BAGS needs more totes to ship frozen products compared to Picnic's current approach and is also more computationally expensive.

The department of Urbanism at the TU Delft, our clients, research the sentiment in different places, times, ages or genders and compare them to each other. This report describes the purpose, design, implementation and accuracy of a web tool created to get insights into the sentiment people have, de- ducted from social media. The aim of our project was to make research easy and extracting innovative insights from social media.

In our tool, we analysed tweets and their location to collect information about the sentiment different people have towards places. The implementation of our tool consisted of five main components: (1) Twitter-Kafka, processing the tweets from the tweet data stream to our database, (2) face recognition, used for determining whether a tweet comes from a person instead of a company or organisation and for age and gender inference, (3) sentiment analysis, using machine learning to determine whether a tweet is neutral, negative or positive, (4) REST API, for the connection between the front-end and the back-end and (5) the user interface, in the form of an interactive dashboard.

At the beginning of the project, we set up a pipeline that checks the code on multiple things. The testing of the back-end is based on a Python unit test suit. For the build to succeed, all tests must pass and the total branch coverage must be at least 80%. We used Flake8 and ESlint in our build to ensure code quality at all times.

All of the above-mentioned components are up and running. The clients are now able to research the sentiments of people towards places.