We started our week with a presentation of the curricular units present this semester, AAUT1IA and PLNTDIA.
AAUT1IA
During the 1st theoretical class of Machine Learning was presented what will be developed in the curricular unit and all the deadlines.
In the 1st practice class, we began to discuss which topic fits better to move forward with the new challenge. We will need to choose a theme and find a DataSet accordingly with the theme.
The themes presented were:
- Music recommendation data set (Spotify) https://research.atspotify.com/datasets/
- Data set of music/podcasts playlists (Spotify) https://research.atspotify.com/datasets/
- Movie recommendations data set (NetFlix) https://data.world/chasewillden/netflix-shows
- Youtube Segments Dataset https://research.google.com/youtube8m/
- Fake news detection https://www.kaggle.com/c/fake-news/data
- Urban Sound Detection https://urbansounddataset.weebly.com/urbansound8k.html
- Automobile Data Set https://archive.ics.uci.edu/ml/datasets/automobile
- Dataset for Automotive Applications https://deepvisualmarketing.github.io/
PLNTDIA:
In the 1st practical class of the week of PLNTDIA, we continued to look for ideas and DataSets to find the most suitable theme for the group.
In the second practice class, we continued to review the Datasets we collected in previous classes. After some clarifications with the teacher, we concluded that most of the datasets we had collected would be either too complex or too out of context. So we decided to focus on one theme and find the ideal Dataset. The theme that we choose was: music (Spotify).
After choosing the theme ‘Music’, we came up with ideas to evaluate whether a song will succeed or not and/or if deserve to be in the Top Charts. The datasets that we found to help us with these ideas were:
- The Spotify Hit Predictor Dataset – https://www.kaggle.com/theoverman/the-spotify-hit-predictor-dataset?select=dataset-of-00s.csv
- Spotify Top 200 Charts (2020-2021) – https://www.kaggle.com/sashankpillai/spotify-top-200-charts-20202021