Python library for multidimensional analysis, classification - clustering analysis and multidimensional analysis
Project description
scientisttools : Exploratory Multivariate Data Analysis with Python
scientisttools is a python package dedicated to multivariate Exploratory Data Analysis, clustering analysis and multidimensional scaling.
Why use scientisttools?
scientisttools provides functions for :
-
Generalized Factor Analysis (GFA) :
- Principal Component Analysis (PCA)
- Principal Component Analysis with partial correlation matrix (PartialPCA)
- Exploratory Factor Analysis (EFA)
- Correspondence Analysis (CA)
- Multiple Correspondence Analysis (MCA)
- Specific Multiple Correspondence Analysis (SpecificMCA)
- Factor Analysis of Mixed Data (FAMD)
- Mixed Principal Component Analysis (MPCA)
- Principal Component Analysis of Mixed Data (PCAMIX)
- Canonical Correlation Analysis (CCA)
- Multiple Factor Analysis (MFA)
- Multiple Factor Analysis for qualitatives/categoricals variables (MFAQUAL)
- Multiple Factor Analysis of Mixed Data (MFAMD)
- Multiple Factor Analysis of Mixed group (MFAMIX)
- Multiple Factor Analysis of Contingence Tables (MFACT)
-
Classification - clustering :
- Hierarchical Clustering on Principal Component (HCPC)
- Variables Hierarchical Clustering Analysis (VARHCA)
- Variables Hierarchical Clustering Analysis on Principal Components (VARHCPC)
- Categorical Variables Hierarchical Clustering Analysis (CATVARHCA)
-
Multidimensional scaling :
- Scaling by MAjorizing a COmplicated Function (SMACOF)
- Classical Multidimensional Scaling (CMDSCALE)
- Metric and Non - Metric Multidimensional Scaling (MDS)
-
In some methods, it allowed to add supplementary informations such as supplementary individuals and/or variables.
-
It provides a geometrical point of view, a lot of graphical outputs.
-
It provides efficient implementations, using a scikit-learn API.
Those statistical methods can be used in two ways :
- as descriptive methods ("datamining approach")
- as reduction methods in scikit-learn pipelines ("machine learning approach")
Notebooks are availabled.
Installation
Dependencies
scientisttools requires
Python >=3.10
numpy >=1.26.4
matplotlib >=3.8.4
scikit-learn >=1.2.2
pandas >=2.2.2
polars >=0.19.2
mapply >=0.1.21
plotnine >=0.10.1
pingouin >=0.5.4
scientistmetrics >=0.0.4
User installation
You can install scientisttools using pip :
pip install scientisttools
Tutorials are available
Examples : Principal component analysis
- Loading data
# Load wine dataset
>>> from scientisttools import load_wine
>>> wine = load_wine()
>>> wine.info()
<class 'pandas.core.frame.DataFrame'>
Index: 21 entries, 2EL to T2
Data columns (total 31 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Label 21 non-null category
1 Soil 21 non-null category
2 Odor.Intensity.before.shaking 21 non-null float64
3 Aroma.quality.before.shaking 21 non-null float64
4 Fruity.before.shaking 21 non-null float64
5 Flower.before.shaking 21 non-null float64
6 Spice.before.shaking 21 non-null float64
7 Visual.intensity 21 non-null float64
8 Nuance 21 non-null float64
9 Surface.feeling 21 non-null float64
10 Odor.Intensity 21 non-null float64
11 Quality.of.odour 21 non-null float64
12 Fruity 21 non-null float64
13 Flower 21 non-null float64
14 Spice 21 non-null float64
15 Plante 21 non-null float64
16 Phenolic 21 non-null float64
17 Aroma.intensity 21 non-null float64
18 Aroma.persistency 21 non-null float64
19 Aroma.quality 21 non-null float64
20 Attack.intensity 21 non-null float64
21 Acidity 21 non-null float64
22 Astringency 21 non-null float64
23 Alcohol 21 non-null float64
24 Balance 21 non-null float64
25 Smooth 21 non-null float64
26 Bitterness 21 non-null float64
27 Intensity 21 non-null float64
28 Harmony 21 non-null float64
29 Overall.quality 21 non-null float64
30 Typical 21 non-null float64
dtypes: category(2), float64(29)
memory usage: 5.3+ KB
- Principal component analysis
>>> from scientisttools import PCA
>>> res_pca = PCA(standardize=True,n_components=5,ind_sup=list(range(15,21)),quanti_sup=[29,30],quali_sup=[0,1],parallelize=True)
>>> res_pca.fit(wine)
- Extract and visualize eigenvalues/varainces:
>>> from scientisttools import get_eig
>>> eig = get_eig(res_pca)
>>> eig.head(6)
eigenvalue difference proportion cumulative
Dim.1 14.740851 11.804968 54.595746 54.595746
Dim.2 2.935884 0.644286 10.873643 65.469389
Dim.3 2.291597 0.569027 8.487397 73.956786
Dim.4 1.722570 0.178368 6.379889 80.336675
Dim.5 1.544202 0.439988 5.719266 86.055941
Dim.6 1.104214 0.355298 4.089681 90.145621
>>> from scientisttools import fviz_screeplot
>>> print(fviz_screeplot(res_pca))
- Extract and visualize results for individuals
>>> # Extract the results for individuals
>>> from scientisttools import get_pca_ind
>>> ind = get_pca_ind(res_pca)
>>> ind.keys()
dict_keys(['coord', 'cos2', 'contrib', 'infos'])
>>> # Factor coordinates for individuals
>>> ind["coord"].head(6)
Dim.1 Dim.2 Dim.3 Dim.4 Dim.5
2EL -0.697060 -2.852968 0.169384 0.942456 0.957457
1CHA -4.802838 -4.236054 -0.969634 0.164637 0.288523
1FON -2.577287 -0.184102 -1.879038 -2.541729 -0.390677
1VAU -10.582720 2.180724 1.691468 -0.624785 0.968510
1DAM 4.906159 0.147573 0.864068 -0.138473 2.525392
2BOU 1.746737 0.926421 -1.281174 -1.113625 -1.805437
>>> # Individuals factor map
>>> from scientisttools import fviz_pca_ind
>>> print(fviz_pca_ind(res_pca))
Author(s)
Duvérier DJIFACK ZEBAZE (djifacklab@gmail.com)
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file scientisttools-0.1.6.tar.gz
.
File metadata
- Download URL: scientisttools-0.1.6.tar.gz
- Upload date:
- Size: 24.8 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.10.5
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 841d3022d41b91be3afe2f6b2d8767e873e3bc0890ecc8b1250ce933ee865692 |
|
MD5 | 5379a07bb7e9b86c0bf68794ed33e84e |
|
BLAKE2b-256 | a6b905f351589e8636eb4aaa2a8d44413668f0a4a07890fde755576277cb3bda |
File details
Details for the file scientisttools-0.1.6-py3-none-any.whl
.
File metadata
- Download URL: scientisttools-0.1.6-py3-none-any.whl
- Upload date:
- Size: 882.4 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/4.0.2 CPython/3.10.5
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 91100746bf47bce41019f57f7cf9701f3f73cbdfe7963ff54583679545f4c643 |
|
MD5 | dd62ee4ed3b659c2a54f74b07cd603d1 |
|
BLAKE2b-256 | 0605497cae5c35c51fbcccefbc13cc6ddb283768daf47a84eba875c3c4f57610 |