[Pieter van der Hijden] - Fab Futures - Data Science

Home About

< Home

DataScience Session 1: Introduction

Synopsis

Neil introduced the Fab Futures program and its first course: Data Science. He presented many examples of infographics and links to works of reference and interesting datasets. Surprise for me was the work by Florence Nightingale. My favourites are the Meinard map of Napoleon in Russian Winter (so much data, presented so clearly and still actual!) and the energy matrix called "US Energy Flow Super Sankey" by Otherlab.

Resources

• The Nature of Mathematical Modeling
• Numerical Recipes
• Jupyter
• data sets
  • awesome public datasets
  • awesome public real-time datasets
  • Kaggle
  • data.gov
  • UC Irvine
  • OpenML
  • AWS
  • sklearn.datasets

My additions:

• Information Graphics
• The Power of Where: A Geographic Approach to the World's Greatest Challenges
• Storytelling with Data: A Data Visualization Guide for Business Professionals
• data sets
  • Fab Network Data
    • Fablab Register
    • Fabcity Members List (informal)
    • Fablab SDG Profiles
  • GeoNames
  • Sustainable Development Report 2025

Assignment

• Select and document a data set to analyze
• Connect to a JupyterLab server and become familiar with the user interface

A. Research ideas

• The dataset I selected is the list of fablabs, called labs.json, as can be downloaded via the https://fablabs.io website.
• For this specific dataset, I document the different fields/columns.
• I apply some techniques to get a first impression of the data; these techniques cann be applied to any dataset.
• I start cleaning this specific dataset and save it to a local CSV file.

B. Research planning and design

• read fablab register and document it
• review and plot the data
• clean the data
• plot some data
  • change empty activity_status to "unknown"
  • select most relevant fields/columns only
  • skip all records for closed fablabs
  • save the dataset to csv

Directories

Manually, we created the following directories:

• outputs (all outputs are written to this directory; this directory has been excluded from the repository)

Common modules

To facilitate the re-use of software, I placed some Python functions in fabmodules.py in the same directory as my Python notebooks. In this way all notebooks can use them. Use "import fabmodules as fm" to activate them.

# import python modules
import pandas as pd
import matplotlib.pyplot as plt
import fabmodules as fm
output_path = fm.output_path = "outputs/"

C. Data collection

Read fablab register as dataframe

The dataset we selected is the fablab register of "today", ie the labs.json export file from https://fablabs.io. We import it into a pandas dataframe.

# read fablab register into a pandas dataframe

url = "https://fablabs.io/labs.json?class=btn+btn-primary"
df = pd.read_json(url)
fm.log("fablabs ",url,df)

2025-12-16T18:14Z fablabs  https://fablabs.io/labs.json?class=btn+btn-primary (2696, 23)

Show dataframe structure

# show dataframe structure and its size
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2696 entries, 0 to 2695
Data columns (total 23 columns):
 #   Column           Non-Null Count  Dtype  
---  ------           --------------  -----  
 0   id               2696 non-null   int64  
 1   name             2696 non-null   object 
 2   kind_name        2696 non-null   object 
 3   parent_id        918 non-null    float64
 4   blurb            2649 non-null   object 
 5   description      2667 non-null   object 
 6   slug             2696 non-null   object 
 7   avatar_url       2179 non-null   object 
 8   header_url       1708 non-null   object 
 9   address_1        2668 non-null   object 
 10  address_2        2666 non-null   object 
 11  city             2696 non-null   object 
 12  county           2696 non-null   object 
 13  postal_code      2694 non-null   object 
 14  country_code     2696 non-null   object 
 15  latitude         2451 non-null   float64
 16  longitude        2451 non-null   float64
 17  address_notes    2667 non-null   object 
 18  phone            2667 non-null   object 
 19  email            2667 non-null   object 
 20  capabilities     2696 non-null   object 
 21  activity_status  2223 non-null   object 
 22  links            2696 non-null   object 
dtypes: float64(3), int64(1), object(19)
memory usage: 484.6+ KB

Data Dictionary

The documentation of our dataset is a so-called data dictionary. We document all the columns (fields) of our dataset. In most cases it is enough to add a meaningful "Label" to eachof them. Where appropriate we give some additional comment.

The following picture shows the data dictionary for our dataset.

D. Data processing

We want to get some first impressions about the data in our dataset. Therefore we use some methods we could use for any dataset. Here we apply them to the fablabs.

Count unique values

We count the unique values in each column and analyze the columns where these unique numbers are max 20. In the fablabs case the columns kind_name has 3 unique values, the column activity_status has 6. For these columns we create barcharts

ChatGPT prompt: pandas: count the unique values in all columns of a dataframe; create a barchart of these figures; how to find the columns with not more than n different values; how to plot separate bar charts for these?

# count unique values in each column
def make_hashable(x):
    if isinstance(x, (list, dict)):
        return str(x)
    return x

unique_counts = df.apply(lambda col: col.map(make_hashable).nunique(dropna=True))


# create bar chart
ax = unique_counts.plot(kind="bar", figsize=(10,5))
fm.label_bars(ax)

plt.title("Number of Unique Values per Column")
plt.xlabel("Column")
plt.ylabel("Unique value count")
plt.xticks(rotation=45, ha="right")

plt.tight_layout()
plt.show()

Create barcharts for columns with max 20 different values

# focus on the columns with less than x different values
maxnum = 20 # max number of different values in column

#✅ Step 1 — Count unique values per column (robust)
def make_hashable(x):
    if isinstance(x, (list, dict)):
        return str(x)
    return x

unique_counts = df.apply(
    lambda col: col.map(make_hashable).nunique(dropna=True)
)
# ✅ Step 2 — Find columns with ≤ 20 unique values
cols_le_maxnum = unique_counts[unique_counts <= maxnum].index.tolist()
print(cols_le_maxnum)

# ✅ Step 3 — Plot separate bar charts for these columns

for col in cols_le_maxnum:
    counts = df[col].map(make_hashable).value_counts(dropna=False)

    plt.figure(figsize=(6, 4))
    ax = counts.plot(kind="bar")
    fm.label_bars(ax)

    plt.title(f"Value distribution: {col}")
    plt.xlabel(col)
    plt.ylabel("Count")
    plt.xticks(rotation=45, ha="right")

    
    plt.tight_layout()
    plt.show()

['kind_name', 'activity_status']

Calculate statistics for numerical columns

We calculate some statistics on numerical columns. In the fablabs case, it concerns four columns. The statistical results are not very meaningfull in this case.

# show some dataframe statistics
df.describe()

	id	parent_id	latitude	longitude
count	2696.000000	918.000000	2451.000000	2451.000000
mean	1569.617953	725.663399	29.624157	-1.136253
std	890.172753	677.779966	24.582534	62.173664
min	1.000000	2.000000	-54.935210	-158.057284
25%	805.750000	166.500000	19.125337	-49.072194
50%	1582.500000	486.000000	39.845776	4.370654
75%	2321.250000	1121.000000	46.335691	24.315401
max	3114.000000	3111.000000	69.576047	175.663262

Display some records (horizontally)

We display a list of records; ie the head and the tail of the list of fablabs.

# list a few records of the dataset
df

	id	name	kind_name	parent_id	blurb	description	slug	avatar_url	header_url	address_1	...	postal_code	country_code	latitude	longitude	address_notes	phone	email	capabilities	activity_status	links
0	2386	Al Jazri Lab	fab_lab	NaN	The Al Jazari Lab (Fabrication Lab) is named i...	The lab is equipped with cutting-edge technolo...	aljazrilab	http://fablabs.io.s3.amazonaws.com/2021/02/06/...	http://fablabs.io.s3.amazonaws.com/2021/02/06/...	House of Wisdom, Al Juraina 1	...		AE	NaN	NaN	Al Jazri Lab is located in Mezzanine Floor at ...	0097165940000	aljazrilab@houseofwisdom.ae	[three_d_printing, cnc_milling, circuit_produc...	active	[{'id': 31737, 'url': 'https://www.facebook.co...
1	17	PiNG	fab_lab	NaN	None	None	ping	None	None	None	...	44000	FR	47.218371	-1.553621	None	None	None	[]	None	[{'id': 52, 'url': 'http://fablab.pingbase.net'}]
2	20	FabLab INSA Strasbourg	fab_lab	NaN	None	None	fablabinsastrasbourg	None	None	None	...	67000	FR	48.583148	7.747882	None	None	None	[]	None	[{'id': 55, 'url': 'http://www.ideaslab.fr'}]
3	2852	FabLab Solidaire Orange Digital Center Côte d'...	fab_lab	2791.0	Le FabLab Solidaire de l’Orange Digital Center...	Basé sur le "faire" et le partage, le FabLab S...	fablabsolidaireorangedigitalcenterctedivoire	http://fablabs.io.s3.amazonaws.com/2024/01/12/...	http://fablabs.io.s3.amazonaws.com/2024/02/09/...	https://maps.app.goo.gl/AYZTpZ8RncKXq7WEA	...	11 BP 202 Abidjan 11	CI	NaN	NaN	Orange Digital Center - Plateau Indénié, 5ème ...	+2252720345555	fablabodcci@gmail.com	[three_d_printing, cnc_milling, laser, vinyl_c...	active	[{'id': 43112, 'url': 'https://www.facebook.co...
4	2963	FabLab Caxias do Sul	fab_lab	NaN	The FabLab Caxias is a Maker Space and a versa...	The Fabrication Laboratory (FabLab) at the Cax...	caxiasfablab	http://fablabs.io.s3.amazonaws.com/2025/03/18/...	http://fablabs.io.s3.amazonaws.com/2024/11/16/...	R. Avelino Antônio de Souza, 1730	...	95043-700	BR	-29.139272	-51.171602		+55 54 9 8116 0397	fablab@caxias.ifrs.edu.br	[three_d_printing, cnc_milling, circuit_produc...	active	[{'id': 61302, 'url': 'https://integra.ifrs.ed...
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2691	13	Fab Lab Egypt	fab_lab	NaN	The first official fab lab and community-run m...	Fab Lab Egypt (under the umbrella of San3a Tec...	fablabegypt	http://fablabs.io.s3.amazonaws.com/2018/03/02/...	http://fablabs.io.s3.amazonaws.com/2021/01/19/...	Villa 35 - 100 St. - Near Al Horia Square	...		EG	29.959218	31.252782		+201017465650	fablabegypt@san3atech.com	[three_d_printing, cnc_milling, circuit_produc...	active	[{'id': 48, 'url': 'https://www.facebook.com/f...
2692	2348	Martigues'FabLab	fab_lab	228.0	Initié par les Espaces Publics Numériques de l...	Présentation en vidéo : https://dai.ly/x87kfi5...	martiguesFabLab	http://fablabs.io.s3.amazonaws.com/2020/11/24/...	http://fablabs.io.s3.amazonaws.com/2021/10/18/...	Quai Lucien Toulmond	...	13500	FR	43.405296	5.051129	EPN Maison de la Formation et de la Jeunesse	0442494598	epn@ville-martigues.fr	[three_d_printing, laser, vinyl_cutting]	active	[]
2693	1730	BISCAST ManFabLab	fab_lab	NaN	Naga's leading center of idea generation and s...	Formed between the partnership of Bicol State ...	biscastmanfablab	http://fablabs.io.s3.amazonaws.com/2018/06/01/...	http://fablabs.io.s3.amazonaws.com/2018/10/08/...	98 Penafrancia Ave.	...	4400	PH	13.636209	123.199228			biscast.manfablab@gmail.com	[three_d_printing, cnc_milling, laser, precisi...	active	[]
2694	1587	The Makerspace (previously TechWorks Amman)	fab_lab	NaN	The Makerspace is CPF’s flagship innovation pl...	The Makerspace (previously TechWorks) is CPF’s...	fablabjordan	http://fablabs.io.s3.amazonaws.com/2024/03/13/...	http://fablabs.io.s3.amazonaws.com/2021/01/20/...	King Hussein Business Park	...		JO	31.971558	35.832745	Next to GROW Building side door entrance	00962791000110	info@techworks.jo	[three_d_printing, cnc_milling, circuit_produc...	active	[{'id': 29176, 'url': 'https://youtu.be/zLt3aO...
2695	2210	FabLab L'Aquila	fab_lab	NaN	FabLabAQ is a group of people hold together by...	FabLab L’Aquila Association\r\n\r\nL'Aquila's ...	fablabaq	http://fablabs.io.s3.amazonaws.com/2020/02/07/...	http://fablabs.io.s3.amazonaws.com/2020/02/07/...	Via Giuseppe Mezzanotte	...	67100	IT	42.354660	13.413750		+393927629526	info@fablaquila.org	[three_d_printing, cnc_milling, circuit_produc...	active	[{'id': 30617, 'url': 'https://www.instagram.c...

2696 rows × 23 columns

Display some records (vertically)

We also can display the records transposed (vertically). Sometimes that is more meaningful.

# list a few records of the dataset TRANSPOSED
df.T

	0	1	2	3	4	5	6	7	8	9	...	2686	2687	2688	2689	2690	2691	2692	2693	2694	2695
id	2386	17	20	2852	2963	790	776	892	2868	1	...	2331	663	214	2336	1108	13	2348	1730	1587	2210
name	Al Jazri Lab	PiNG	FabLab INSA Strasbourg	FabLab Solidaire Orange Digital Center Côte d'...	FabLab Caxias do Sul	Drone Lab Brasil	Fab Lab Lleida	FABLAB KERALA - Trivandrum	Ostriv Platform	Fab Lab Afghanistan	...	FabLab_du_Rocher	Confluence Fablab	Fablab Tainan	Konk Ar Lab	Fablab STMC	Fab Lab Egypt	Martigues'FabLab	BISCAST ManFabLab	The Makerspace (previously TechWorks Amman)	FabLab L'Aquila
kind_name	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	mini_fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	...	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab	fab_lab
parent_id	NaN	NaN	NaN	2791.0	NaN	NaN	NaN	NaN	1490.0	NaN	...	NaN	NaN	NaN	1087.0	214.0	NaN	228.0	NaN	NaN	NaN
blurb	The Al Jazari Lab (Fabrication Lab) is named i...	None	None	Le FabLab Solidaire de l’Orange Digital Center...	The FabLab Caxias is a Maker Space and a versa...	Laboratory specialized in using the Drone as a...	Help the community around us to develop social...	With the support of Government of Kerala, KSUM...	Welcome to Ostriv Platform! Our space is divid...	None	...	A FabLab in the center of Paris (France), dedi...	Full CAD suite of software (Autodesk, Solidwor...	Fab Lab Tainan was founded at Dec. 2013 at Tai...	Konk Ar Lab, le FabLab de la Baie, est un FabL...	Fablab STMC（Southern Taiwan Maker Center）found...	The first official fab lab and community-run m...	Initié par les Espaces Publics Numériques de l...	Naga's leading center of idea generation and s...	The Makerspace is CPF’s flagship innovation pl...	FabLabAQ is a group of people hold together by...
description	The lab is equipped with cutting-edge technolo...	None	None	Basé sur le "faire" et le partage, le FabLab S...	The Fabrication Laboratory (FabLab) at the Cax...	Drone Lab Brasil was born as part of the Miran...	We are looking for a location for our lab.	With the support of Government of Kerala, KSUM...	Welcome to Ostriv Platform! Our space is divid...	None	...	A FabLab in the center of Paris (France), dedi...	Confluence fablab is located just outside St. ...	We are Fablab Tainan, one of the main Fablabs ...	Konk Ar Lab est une association loi 1901 créée...	As a response to the concept of “social design...	Fab Lab Egypt (under the umbrella of San3a Tec...	Présentation en vidéo : https://dai.ly/x87kfi5...	Formed between the partnership of Bicol State ...	The Makerspace (previously TechWorks) is CPF’s...	FabLab L’Aquila Association\r\n\r\nL'Aquila's ...
slug	aljazrilab	ping	fablabinsastrasbourg	fablabsolidaireorangedigitalcenterctedivoire	caxiasfablab	dronelabbrasil	fablablleida	fablabkeralatrivandrum	ostrivplatform	fablabafghanistan	...	fablabdurocher	confluencefablab	fablabtainan	konkarlab	fablabstmc	fablabegypt	martiguesFabLab	biscastmanfablab	fablabjordan	fablabaq
avatar_url	http://fablabs.io.s3.amazonaws.com/2021/02/06/...	None	None	http://fablabs.io.s3.amazonaws.com/2024/01/12/...	http://fablabs.io.s3.amazonaws.com/2025/03/18/...	http://fablabs.io.s3.amazonaws.com/2020/05/25/...	http://fablabs.io.s3.amazonaws.com/2017/01/28/...	http://fablabs.io.s3.amazonaws.com/2018/08/01/...	None	None	...	http://fablabs.io.s3.amazonaws.com/2020/11/01/...	http://fablabs.io.s3.amazonaws.com/2017/01/28/...	http://fablabs.io.s3.amazonaws.com/2017/01/28/...	http://fablabs.io.s3.amazonaws.com/2020/11/09/...	http://fablabs.io.s3.amazonaws.com/2017/01/28/...	http://fablabs.io.s3.amazonaws.com/2018/03/02/...	http://fablabs.io.s3.amazonaws.com/2020/11/24/...	http://fablabs.io.s3.amazonaws.com/2018/06/01/...	http://fablabs.io.s3.amazonaws.com/2024/03/13/...	http://fablabs.io.s3.amazonaws.com/2020/02/07/...
header_url	http://fablabs.io.s3.amazonaws.com/2021/02/06/...	None	None	http://fablabs.io.s3.amazonaws.com/2024/02/09/...	http://fablabs.io.s3.amazonaws.com/2024/11/16/...	http://fablabs.io.s3.amazonaws.com/2020/05/25/...	None	http://fablabs.io.s3.amazonaws.com/2020/11/05/...	None	None	...	http://fablabs.io.s3.amazonaws.com/2020/11/01/...	http://fablabs.io.s3.amazonaws.com/2017/01/28/...	http://fablabs.io.s3.amazonaws.com/2021/01/01/...	http://fablabs.io.s3.amazonaws.com/2021/01/18/...	http://fablabs.io.s3.amazonaws.com/2017/01/28/...	http://fablabs.io.s3.amazonaws.com/2021/01/19/...	http://fablabs.io.s3.amazonaws.com/2021/10/18/...	http://fablabs.io.s3.amazonaws.com/2018/10/08/...	http://fablabs.io.s3.amazonaws.com/2021/01/20/...	http://fablabs.io.s3.amazonaws.com/2020/02/07/...
address_1	House of Wisdom, Al Juraina 1	None	None	https://maps.app.goo.gl/AYZTpZ8RncKXq7WEA	R. Avelino Antônio de Souza, 1730	Avenida Brigadeiro luis Antonio, 580		Kerala Startup Mission (KSUM)	Volynska Street, 9/21, Kyiv, Україна	Jalalabad, Afghanistan	...	230 Rue du Faubourg Saint-Honoré	NO Nelson - N7	No. 10-1, Ln. 197, Yule St., East Dist.	2 Rue des Charmes	No.32, Datong Rd.	Villa 35 - 100 St. - Near Al Horia Square	Quai Lucien Toulmond	98 Penafrancia Ave.	King Hussein Business Park	Via Giuseppe Mezzanotte
address_2		None	None			5º andar		Ground floor IIITM-K building , Technopark Campus		None	...			B2	Dans l'enceinte de l'école primaire	Sinying Dist.	Maadi		1st floor, College of Engineering and Architec...	Building 23	C/O IIS "Amedeo D'Aosta
city		Nantes	Strasbourg	Abidjan	Caxias do Sul	São Paulo	Lérida	Trivandrum	Kyiv	Jalalabad	...	Paris	Edwardsville	Tainan City	Concarneau	Tainan City	Cairo	Martigues	Naga City	Amman	L'Aquila
county	Sharjah	Pays de la Loire	Alsace	Abidjan/Plateau/Indénié	Rio Grande do Sul	São Paulo	Lleida	KERALA	Ukraine	Nangarhar	...		Illinois	Taiwan	Bretagne		Cairo	Bouches du Rhône	Camarines Sur	Jordan	L'Aquila
postal_code		44000	67000	11 BP 202 Abidjan 11	95043-700	01318-000		699581	03151	None	...	75008	62025	701	29900	73048		13500	4400		67100
country_code	AE	FR	FR	CI	BR	BR	ES	IN	UA	AF	...	FR	US	TW	FR	TW	EG	FR	PH	JO	IT
latitude	NaN	47.218371	48.583148	NaN	-29.139272	-23.542037	41.623023	8.557167	50.430186	34.434469	...	48.876448	38.806278	22.993373	47.88356	23.130636	29.959218	43.405296	13.636209	31.971558	42.35466
longitude	NaN	-1.553621	7.747882	NaN	-51.171602	-46.6358	0.625864	76.880297	30.451312	70.448505	...	2.303581	-89.949941	120.215284	-3.915601	120.290958	31.252782	5.051129	123.199228	35.832745	13.41375
address_notes	Al Jazri Lab is located in Mezzanine Floor at ...	None	None	Orange Digital Center - Plateau Indénié, 5ème ...		perto do metro Sé				None	...		The lab must be entered through the Southeast ...					EPN Maison de la Formation et de la Jeunesse		Next to GROW Building side door entrance
phone	0097165940000	None	None	+2252720345555	+55 54 9 8116 0397	+5511971835193	+34 660493352	+91 9809494669	tel:+380631537951	None	...	+33180984657	618 468 5941	886-6-2212245	0256100304	886-6-6356583	+201017465650	0442494598		00962791000110	+393927629526
email	aljazrilab@houseofwisdom.ae	None	None	fablabodcci@gmail.com	fablab@caxias.ifrs.edu.br	contato@dronelab.com.br	fablablleida@gmail.com	fablab@startupmission.in	mailto:ostrivplatform@gmail.com	None	...	fablab@coursdurocher.fr	info.stlcfl@lc.edu	fablabtainan@gmail.com	contact@konkarlab.bzh	stmakercenter@gmail.com	fablabegypt@san3atech.com	epn@ville-martigues.fr	biscast.manfablab@gmail.com	info@techworks.jo	info@fablaquila.org
capabilities	[three_d_printing, cnc_milling, circuit_produc...	[]	[]	[three_d_printing, cnc_milling, laser, vinyl_c...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, circuit_production, laser, ...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, cnc_milling, circuit_produc...	[]	...	[three_d_printing, laser, vinyl_cutting]	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, cnc_milling, laser, vinyl_c...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, laser, vinyl_cutting]	[three_d_printing, cnc_milling, laser, precisi...	[three_d_printing, cnc_milling, circuit_produc...	[three_d_printing, cnc_milling, circuit_produc...
activity_status	active	None	None	active	active	active	planned	active	active	None	...	active	active	active	active	None	active	active	active	active	active
links	[{'id': 31737, 'url': 'https://www.facebook.co...	[{'id': 52, 'url': 'http://fablab.pingbase.net'}]	[{'id': 55, 'url': 'http://www.ideaslab.fr'}]	[{'id': 43112, 'url': 'https://www.facebook.co...	[{'id': 61302, 'url': 'https://integra.ifrs.ed...	[{'id': 11077, 'url': 'https://www.instagram.c...	[{'id': 2145, 'url': 'http://www.fablablleida....	[{'id': 1563, 'url': 'http://www.startupmissio...	[{'id': 43268, 'url': 'https://ostrivplatform....	[{'id': 179, 'url': 'http://www.fablab.af'}]	...	[{'id': 19805, 'url': 'http://www.petitcoursdu...	[]	[{'id': 2131, 'url': 'http://www.fablabtainan....	[{'id': 20556, 'url': 'https://www.konkarlab.b...	[{'id': 1968, 'url': 'http://southmaker.wda.go...	[{'id': 48, 'url': 'https://www.facebook.com/f...	[]	[]	[{'id': 29176, 'url': 'https://youtu.be/zLt3aO...	[{'id': 30617, 'url': 'https://www.instagram.c...

23 rows × 2696 columns

E. Data Study and Analysis

In this case we are replacing empty values, removing (from our dataset) the fablabs that have been closed, and skipping the fields we arenot interested in anymore.

Replace empty activity_status with "unknown"

# maybe some values are empty; replace them by "unknown"
df["activity_status"] = df["activity_status"].replace("","unknown")
df['activity_status'] = df['activity_status'].fillna('unknown')

Remove records with activity_status == "closed"

# remove records you are not interested in; eg remove closed fablabs
print(df.shape)
df = df[df["activity_status"]!="closed"]
print(df.shape)

(2696, 23)
(2603, 23)

Remove any fields you are not interested in

# use only fields you are interested in
print(df.shape)
columnlist = "slug,name,kind_name,activity_status,country_code,city".split(",")
df = df[columnlist]
print(df.shape)
df.info()

(2603, 23)
(2603, 6)
<class 'pandas.core.frame.DataFrame'>
Index: 2603 entries, 0 to 2695
Data columns (total 6 columns):
 #   Column           Non-Null Count  Dtype 
---  ------           --------------  ----- 
 0   slug             2603 non-null   object
 1   name             2603 non-null   object
 2   kind_name        2603 non-null   object
 3   activity_status  2603 non-null   object
 4   country_code     2603 non-null   object
 5   city             2603 non-null   object
dtypes: object(6)
memory usage: 142.4+ KB

F. Data Publishing and Access

Make a barchart and save picture

fm.plotbarchart(df,"activity_status")

G. Data Preservation

Save the dataframe

# save the resulting dataframe for future use
df.to_csv(output_path+"dataframe.csv", index = False)
df.shape

(2603, 6)

H. Data Re-use

Read the saved dataframe

# at any time you can read this file back into pandas
df1 = pd.read_csv(output_path+"dataframe.csv")
df1.shape

(2603, 6)

Create a combined fabnetdata.xlsx

Use the MONTHLY - Create FABNETDATA workbook to create an excel workbook with separate sheets for fablabs, fabcities, countries and continents, plus their data dictionaries. The Notebook combines data collection, data cleaning and running statistics. The WG Fablabs and SDGs runs it on the first day of every calendar month and stores the resulting fabnetdata.xlsx in the public folder of the repository. However, you can also run the notebook yourself at any time you want.

Read the combined fabnetdata.xlsx

Use the read_fabnetdata module in fabmodules.py to read an existing fabnetdata.xlsx file and disassemble it into dataframes fablabs, fabcities, countries, continents; it clearly states the original data retrieval date. TRY!

# Read fabnetdata.xlsx
# read fabnetdata
fablabs, fabcities, countries, continents, retrieved = fm.read_fabnetdata("fabnetdata.xlsx")

url https://gitlab.fabcloud.org/fl-management/fablab-network-data/-/raw/main/public/fabnetdata.xlsx
fablabs  (2597, 29)
fabcities  (56, 7)
countries  (250, 44)
continents  (7, 35)
data retrieved 2025-12-01 

Evaluation and Follow-up

I used ChatGPT for adhoc assistance with smaller coding issues. The software for barplots was a collaborative effort.

Follow-up

• Solve the missing "NA" strings issue when exporting a dataset to CSV and importing it again ("NA"returns as NaN).
• Create a python function for a quick and dirty first review of a dataset (a combi of the activities practices above).
• Figure out how the fabmodules.py module could be made avalable outside the current directory (Jupyter requires a copy in the same directory as the calling notebook).

Review