This course is part of the Scientific Data Analytics and Modelling Programme. The aim of the course is that students gain practical skills in accessing large databases/datasets, handling data stored in different formats, exploring/cleaning these data and presenting the gathered information. During the course students will come across databases used in various scientific fields. Completing of the several projects allows students to gain useful experiences that will give a firm foundation for later courses on theoretical datamining and advanced computing laboratories. In this course we intend to introduce state of the art tools and methods for data exploration and visualization. This field evolves rapidly and therefore the emphasis is not on the exact tool for solving a task, but the notions and algorithms. A library in a given programming language that we use today might become obsolete in a year or become the standard for a longer time.
There is a useful tutorial into python http://bokae.web.elte.hu/numerical_methods.html), which gives a wide background knowledge, that comes handy.
During the course every sample code will be shown in jupyter notebooks, which can be accessed on the K8plex Edu platform.
Each occasion starts with approximately an hour of introduction into the current topic with examples. After that everyone can work on the assignments and the lecturers will be available to help with the any related problems and questions.
Neptun code: dsexplorf20vm
Instructor: Dávid Visontai
Semester: spring
Type: Lecture + Practice
Credit points: 6
Prerequisites: programming in either python, R (or matlab)
The course is held in the North Building in computer lab 5.56 or online in Teams every Wednesday between 2PM-4PM. Attendance on the lectures is not compulsory but it is highly recommended, especially when we can meet personally again because of the consultation. There are many small problems that one might encounter and can be easily explained personally.
1, 09.02.2021. Introduction to K8plex Edu, Jupyter Notebooks, Requirements for the course
2, 16.02.2022. Timeseries analysis and USGS water discharge statistics
3, 23.02.2022. Maps, shapes, coordinates
4, 02.03.2022. Interactive Visualization
5, 09.03.2022. REST services
6, 16.03.2022. SQL queries
7, 23.03.2022. Datacleaning, and filtering
8, 30.03.2022. Network exploration - This lecture will be given by Dániel Ábel, who is a developer at Maven7.
9, 06.04.2022. Natural Language Processing on tweets
10, 20.04.2022. Consultation only
11, 27.04.2022. Image exploration
12, 04.05.2022. Working with large datasets
13, 11.05.2022. **Consultation only **
- Datatypes, images, timeseries, tables, graphs, textual data
- Standards of file- and dataformats (csv, hdf5, netcdf)
- Raw and processed data, metadata, cleansing of data
- Access data locally and through the web (APIs, HTTP protocol)
- Access of scientific databases, Usage of relational databases (SQL)
- Transforming data, sorting, combining pandas
- Basics of timeseries analysis
- Handling datasets with geolocation (shapely, folium, geoviews)
- Basics of image processing (opencv)
- Dimension reduction, clustering
- Processing textual data, logs (natural language processing)
- Infographics, visualisation (html, css, javascript libraries)
- Interactive dataexplorative tools (ipywidgets, bokeh, holoviews)
- Developing open source softwares, reproducible research (OSF)
- Following John Snow - HTML - Ágnes Becsei
- USGS water discharge statistics - HTML - Dávid Visontai
- SQL queries on an NBA database- HTML - Dávid Visontai
- Interactive Visualization - HTML or Hosted App - Krisztián Papp
- REST API - (REST service/API) a notebook or a python script - Dávid Visontai
- Network exploration - HTML - István Márkusz
- Data extraction from images - HTML - József Stéger
- Natural Language Processing on tweets - HTML - !!! Ex 7. and 8. is optional. - István Márkusz
All assigments should be converted into HTML unless it is stated otherwise!
The submitted results should look like a report, which can be generated from the notebooks. This will be explained on the lecture. All figures should have labels, title, each exercise should end with a descriptive conclusion and explanatory comments should be inserted into the code. These are must have features of a work that is intended for presentation.
The assignments/projects will be released in the first couple of weeks, which you have to submit after the end of the course 16th June 1PM (this is the last deadline). Each submitted assignment will be corrected and a feedback will be given. After that you may resubmit your assignment once more but before the final deadline.
https://k8plex-edu.elte.hu/hub is where the notebooks will be handed out. It is available for all students with a valid Neptun or caesar account. Once you run your notebook server you will find a folder with the course material. The notebooks will be available in this Github repository as well. We will explain how to use this portal on the first lecture. The K8plex Edu platform is accessible externally as well. In case there is any problem with the portal you can run a notebook server locally on any other computer and upload your work later.
Each assignment will be corrected after submission and a maximum of 20 points will be given for it. 10 for all the completed tasks, 10 for the quality of the submitted report (look, clarity and comments).
The final grades will be given according to the following point system:
0 - 50: failed
51 - 70: 2
71 - 94: 3
95 - 115: 4
116 - : 5
- Assignments will be visible in
/v/courses/dataexpviz.assignments - Example Datasets will be in the
/v/courses/dataexpviz.public/Datasetsdirectory you will find datasets, that you can work with. - Other lecture materials will be in
/v/courses/dataexpviz.public/<lecture_name>and on this github page
- Python tutorial: http://bokae.web.elte.hu/numerical_methods.html (translated from the BSc course "numerical methods in physics I" by Eszter Bokányi, work in progress )
- SQL tutorial: https://www.w3schools.com/sql/
- RESTful service: https://en.wikipedia.org/wiki/Representational_state_transfer
- Networkx: https://networkx.github.io/
- Wes McKinney: Python for Data Analysis, (O’Reilly 2013)
- Joel Grus: Data Science from Scratch (O’Reilly 2015)
- Ágnes Becsei
- Krisztián Papp
- István Márkusz
- József Stéger
- Dávid Visontai