COS80023 - Topics suitable for D/HD projects

These topics are by no means exclusive. You may have a technology of your own you want to investigate. The best way to avoid nasty surprises is to contact the convenor and discuss intended projects.

The following is a listing of possible topics that are suited for D/HD level work. It is expected that students on this level work largely independently and apply problem-solving skills that are characteristic of IT work.

A project must not be a literature study only, where web pages are summarised and collected into a document. It must include some exploratory work.

Project topic D level (example) HD level (example)
Explore NoSQL database:MongoDB, Cassandra, CouchDB, Neo4J Investigate use (how to create a database, add data in suitable format, query data in various ways) Investigate performance or advanced features
Explore Hadoop Technologies – including ones already introduced in the P/C level work:Hive, Pig, Spark, Kafka, Storm (the latter

three are for good programmers)

Explore the usage and demonstrate the usefulness of a tool for a certain task Compare two tools, or provide advanced analysis, performance analysis or similar
Jupyter notebook Explore functionalities, a few visualisations Explore advanced functionalities or use to analyse complex dataset
Python tools Use python to analyse data, e.g. clustering, classification Carry out more demanding analyses
Deep Learning: Tensorflow, Theano, Torch, Python, Skala (most of the latterrequire programming) Basic neural networks More demanding set- ups and good analyses
Analysis of data sets (e.g. from Kaggle: www.kaggle.com), or data set on mobile counting (ask convenor) Depends on project; confirm with convenor Depends on project; confirm with convenor
Sentiment analysis in natural language processing Application of simple tools, e.g. in R with reasonably good outcomes Application of simple tools, e.g. in R with excellent insights
Long/short term traffic prediction. Data set provided. Can use neural networks or arima models. Choice of tool should be argued. Basic outcomes or simple design of model, but generally good approach. Well argued choice of tool, good outcomes, suitable analysis.

Expert's Answer

help

Hire Expert 

Get a Professional Help


200
Select FileChangeRemove

TOP