pyethnobiology 0.1.1

Ethnobiological data assessment tool.

pyethnobiology

Ethnobiological data assessment tool

Introduction

This tool is designed to help researchers analyze their ethnobiological data. It is based on the ethnobotanyR package with some modifications. Also, for further information on the methods used in this tool, please refer to the ethnobotanyR package.

Scientific explanations of the indices used in this tool can be found in the documentation of the ethnobotanyR package.

Installation

To install this package, you can use the following code:

pip install pyethnobiology

We recommend use virtualenv to install this package.

Getting started

Importing the package

import pandas as pd
from pyethnobiology import pyethnobiology

Accepted data formats

This package accepts data in the following formats:

• A data frame with the following columns: informant, taxon, uses. The use columns must be binary (0 or 1). For example:

informant	sp_name	Use_1	Use_2	Use_3	Use_4	Use_5
inform_a	sp_a	0	0	0	1	1
inform_a	sp_b	0	0	0	0	0
inform_a	sp_c	0	0	1	0	0
inform_a	sp_d	0	0	0	0	0
inform_b	sp_a	0	1	0	0	1

• A data frame with the following columns: informant, taxon, use. The use column must be a string. For example:

informant	sp_name	ailments_treated
inform_a	sp_a	Use_3
inform_a	sp_a	Use_8
inform_a	sp_a	Use_9
inform_a	sp_c	Use_6
inform_a	sp_c	Use_7
inform_b	sp_d	Use_3
inform_b	sp_d	Use_6
inform_b	sp_d	Use_7
inform_b	sp_d	Use_8

• Literature data is optional for Jaccard index calculation. If there is a multiple literatures, they must be separated by a semicolon. For example:

informant	sp_name	ailments_treated	literature
inform_a	sp_a	Use_3	1
inform_a	sp_a	Use_8	1;2
inform_a	sp_a	Use_9	1;3;5
inform_a	sp_c	Use_6	2
inform_a	sp_c	Use_7	1;7
inform_b	sp_d	Use_3	3

To use this package, you can use the following code:

Load data from csv, excel or txt file

# You can define your separator, for example: sep=";" or sep="\t". By default, the separator is ",".

data = pd.read_csv("{path}/data.csv") # or pd.read_excel("data.xlsx") or pd.read_csv("data.txt")

Load rda (R Data) file

data = "{path}/data.rda"

Define the columns

If uses are binary values in your data (first example in accepted data formats), you don't need to define the use column. But if uses are strings (second example in accepted data formats), you have to define the use column. For example:

# if uses are binary values

informant_column = "informant"
taxon_column = "sp_name"
convert_use_data = True

# if uses are strings

informant_column = "informant"
taxon_column = "sp_name"
use_column = "ailments_treated"

# if you have literature data

informant_column = "informant"
taxon_column = "sp_name"
use_column = "ailments_treated"
literature_column = "literature"

Create an instance of the class

pye = pyethnobiology(data, informant_column, taxon_column, convert_use_data)

You can also define the use column and literature column in the class instance. For example:

pye = pyethnobiology(data, informant_column="informant", taxon_column="sp_name", use_column="ailments_treated")

Indice calculations

Use Report (UR) per species

pye.UR().calculate()

Out[1]: 
  sp_name  UR
0    sp_c  52
1    sp_a  43
2    sp_d  43
3    sp_b  36

Cultural Importance (CI) index

pye.CI().calculate()

Out[1]: 
  sp_name  CI
0    sp_c  2.60
1    sp_a  2.15
2    sp_d  2.15
3    sp_b  1.80

Frequency of Citation (FC) per species

pye.FC().calculate()

Out[1]: 
  sp_name  FC
0    sp_c  17
1    sp_a  15
2    sp_b  12
3    sp_d  12

Number of Uses (NU) per species

pye.NU().calculate()

Out[1]: 
  sp_name  NU
0    sp_c   8
1    sp_d   8
2    sp_a   7
3    sp_b   7

Relative Frequency of Citation (RFC) index

pye.RFC().calculate()

Out[1]:
  sp_name   RFC
0    sp_c  0.85
1    sp_a  0.75
2    sp_b  0.60
3    sp_d  0.60

Relative Importance (RI) index

pye.RI().calculate()

Out[1]:
  sp_name  RI
0    sp_c  1.000000
1    sp_a  0.878676
2    sp_d  0.852941
3    sp_b  0.790441

Use Value (UV) index

pye.UV().calculate()

Out[1]:
  sp_name    UV
0    sp_c  2.60
1    sp_a  2.15
2    sp_d  2.15
3    sp_b  1.80

Cultural Value (CV) for ethnospecies

pye.CV().calculate()

Out[1]:
  sp_name       CV
0    sp_c  1.76800
2    sp_a  1.12875
1    sp_d  1.03200
3    sp_b  0.75600

Fidelity Level (FL) per species

pye.FL().calculate()

Out[1]:
sp_name    ailments_treated         FL
1     sp_a           Use_10  20.000000
2     sp_a            Use_2  53.333333
3     sp_a            Use_3  60.000000
5     sp_a            Use_5  33.333333
6     sp_a            Use_6  46.666667
8     sp_a            Use_8  40.000000
9     sp_a            Use_9  33.333333
10    sp_b            Use_1  50.000000
13    sp_b            Use_3  41.666667
14    sp_b            Use_4  33.333333
15    sp_b            Use_5  33.333333
16    sp_b            Use_6  41.666667
17    sp_b            Use_7  58.333333
19    sp_b            Use_9  41.666667
20    sp_c            Use_1  52.941176
21    sp_c           Use_10  35.294118
22    sp_c            Use_2  41.176471
24    sp_c            Use_4  35.294118
25    sp_c            Use_5  17.647059
26    sp_c            Use_6  23.529412
27    sp_c            Use_7  29.411765
28    sp_c            Use_8  70.588235
30    sp_d            Use_1  41.666667
32    sp_d            Use_2  25.000000
33    sp_d            Use_3  58.333333
35    sp_d            Use_5   8.333333
36    sp_d            Use_6  41.666667
37    sp_d            Use_7  66.666667
38    sp_d            Use_8  75.000000
39    sp_d            Use_9  41.666667

Informant Consensus Factor (FIC)

Informant Consensus Factor (FIC) operates like a consensus barometer, assessing the agreement among individuals. It boils down to a simple equation:

where Nur is the number of use reports for a particular use category and Nt is the number of taxa used for that category. The FIC ranges from 0 to 1, with 0 indicating no agreement among informants and 1 indicating that all informants agree on the taxa to be used for a particular category (Heinrich et al., 1998; Güzel et al., 2015).

pye.FIC().calculate()

Out[1]:
  ailments_treated       FIC
0            Use_8  0.923077
1            Use_3  0.900000
2            Use_7  0.894737
3            Use_1  0.894737
4            Use_4  0.888889
5            Use_2  0.882353
6           Use_10  0.875000
7            Use_9  0.857143
8            Use_6  0.850000
9            Use_5  0.750000

Jaccard similarity index

Jaccard similarity index is a measure of similarity between two sets of data. The higher the Jaccard index, the more similar the two samples are. You can calculate the Jaccard index for all studies or for each study. For example:

pye.jaccard()

Out[1]:
  study  similarity
0     2    0.480000
1     3    0.480000
2     5    0.480000
3     6    0.440000
4     1    0.400000
5     4    0.360000
6     7    0.360000
7     9    0.200000
8    10    0.035714
9     8    0.000000

Mean similarity of all studies:

pye.jaccard()['similarity'].mean()

Out[1]:
    0.32357142857142857

# Round to 2 decimal places
pye.jaccard()['similarity'].mean().round(2)

Out[1]:
    0.32

Visualization

Radial plot

You can plot all indices in a radial plot except Fidelity Level (FL). For example:

#Use Report (UR) per species
pye.UR().plot_radial()

Customization of the radial plot

You can customize the radial plot. For example:

#Use Report (UR) per species
pye.UR().plot_radial(filename="UR.png", # You can define the saving filename. It is default to "UR.png" for UR index.
                     dpi=300, # You can define the dpi. It is default to 300.
                     num_row=10, # It is default to 10 which is the first 10 rows of the data.
                     ytick_position="onbar", # You can define the ytick position. It is default to "onbar".
                     colors=None, # You can define the colors. It is default to None.
                     show_colorbar=True # You can define if you want to show the colorbar. It is default to True.
                     )

num_row means the number of species you want to plot. For example, if you want to plot the first 5 species of the data, you can use the following code:

#Use Report (UR) per species
pye.UR().plot_radial(num_row=5)

You can change the colors of the radial plot. It takes the range of two colors. For example:

#Use Report (UR) per species
pye.UR().plot_radial(colors=["#ECF4D6", "#265073"]) # [minimum value color, maximum value color]

You can also use multiple configurations. For example:

#Use Report (UR) per species
pye.UR().plot_radial(filename="UR_conf.png",
                     dpi=300,
                     num_row=5,
                     ytick_position="onbar", # on_line
                     colors=["#ECF4D6", "#265073"],
                     show_colorbar=False 
                     )

Heatmap

You can only plot the Fidelity Level (FL) index in a heatmap. For example:

#Fidelity Level (FL) per species
pye.FL().plot_heatmap()

Customization of the heatmap

You can customize the heatmap. For example:

#Fidelity Level (FL) per species
pye.FL().plot_heatmap(
                     filename="FL.png", # You can define the saving filename. It is default to "FL.png"
                     cmap="coolwarm", # You can define the cmap. It is default to "coolwarm".
                     show_colorbar=True, # You can define if you want to show the colorbar. It is default to True.
                     colorbar_shrink=0.50, # You can define the colorbar shrink. It is default to 0.50.
                     plot_width=10, # You can define the plot width. It is default to 10.
                     plot_height=8, # You can define the plot height. It is default to 8.
                     dpi=300, # You can define the dpi. It is default to 300.
                     fillna_zero=True # You can define if you want to fill the NaN values with 0. It is default to True.
                     )
)

Supported cmaps: https://matplotlib.org/stable/tutorials/colors/colormaps.html

#Fidelity Level (FL) per species
pye.FL().plot_heatmap(filename="FL_reds.png",cmap="Reds")

Chord plot

You can plot all data with the Chord plot. It is diagram of ethnobiological uses and species. For example:

pye.plot_chord()

Customization of the Chord plot

You can customize the Chord plot. For example:

pye.plot_chord(filename="chord_plot.png", # You can define the saving filename. It is default to "chord_plot.png"
               dpi=300, # You can define the dpi. It is default to 300.
               by="taxon", # You can define the by. It is default to "taxon". Other option is "informant".
               colors=None, # You can define the colors. It is default to None.
               min_info_count=0, # You can define the minimum information count for filter data. It is default to 0.
               get_first=None # You can define the number of first data. It is default to None.
               )

You can change the colors of the Chord plot. Supported cmaps: https://matplotlib.org/stable/tutorials/colors/colormaps.html

pye.plot_chord(colors="Dark2")

You can also use multiple configurations. For example:

pye.plot_chord(filename="chord_plot_conf.png",
               dpi=300,
               by="informant",
               colors="tab20b",
               min_info_count=5,
               get_first=10
               )