Streamlining Data Extraction in Healthcare with Python in 2023

In recent years, the use of Python in healthcare has increased dramatically. Python is a universal language that can be used for a wide variety of tasks, including data extraction.

With many healthcare software available, data extraction is a critical part of this industry, as it allows researchers to access the data they need to make discoveries and develop new treatments. Python is an ideal language for data extraction due to its flexibility and ease of use.

In this blog post, we will explore how Python can be used to streamline data extraction in healthcare. We will discuss the benefits of using Python for data extraction and how Python can be used to automate repetitive tasks. We will also provide a few examples of how Python can be used to extract data from healthcare databases.

I know that the data is pretty clear, who spends the least and who spends the most but I wanted to take the idea further using this table. I had been looking for a chance to practice web scraping and Visualization in Python and decided this was a great short project.

Although it almost certainly would have been faster to manually enter the data in Excel, then I would not have had the invaluable opportunity to practice a few skills! Data science is about solving problems using a diverse set of tools, and web scraping and regular expressions are two areas I need some work on (not to mention that making plots is always fun). The result was a very short — but complete — project showing how we can bring together these three techniques to solve a data science problem.

Why Python is the Ideal Language for Data Extraction in Healthcare

Python is an ideal language for data extraction in healthcare due to its simple syntax and powerful data structures. Python’s “batteries included” approach makes it easy to install and use external libraries for data analysis, such as NumPy and pandas.

Additionally, Python’s rich set of built-in libraries, such as the os and subprocess modules, make it easy to interface with external programs and systems.

The Various Ways Python Can Be Used for Data Extraction in Healthcare

Python can be used for data extraction in healthcare in a number of ways. One way is through the use of Python libraries such as the Python Imaging Library (PIL) or the PythonMagick library. These libraries can be used to extract data from images, such as medical images. Another way is through the use of Python modules such as the os and shutil modules. These modules can be used to extract data from files or directories.

Requirements

Generally, web scraping is divided into two parts:

1. Fetching data by making an HTTP request
2. Extracting important data by parsing the HTML DOM

Libraries & Tools

1. Beautiful Soup is a Python library for pulling data out of HTML and XML files.
2. Requests allow you to send HTTP requests very easily.
3. Web Scraper will help us to scrape dynamic websites without setting up any automation browser.
4. Pandas is a Python package providing fast, flexible, and expressive data structures
5. matplotlib is a comprehensive library for creating static, animated, and interactive visualizations in Python.

Setup

Our setup is pretty simple. Just create a folder and install Beautiful Soup & Requests. For creating a folder and installing libraries type the below given commands. I am assuming that you have already installed Python 3.x.

mkdir scraper
pip install beautifulsoup4
pip install requests
pip install matplotlib
pip install pandas

Now, create a file inside that folder by any name you like. I am using scraping.py. Then just import Beautiful Soup & requests in your file shown below.

import pandas as pd
from bs4 import BeautifulSoup
import matplotlib.pyplot as plt
import requests

What we are going to scrape

1. Name of the country
2. Expense per capita

Web Scraping

Now, since we have all the ingredients to prepare the scraper, we should make a GET request to the target URL to get the raw HTML data.

r = requests.get('<a href="https://api.scrapingdog.com/scrape?api_key=5ea317fee83f583332581a67&amp;url=https://data.worldbank.org/indicator/SH.XPD.CHEX.PC.CD?most_recent_value_desc=false&amp;dynamic=true%27).text" rel="noreferrer noopener" target="_blank">https://api.scrapingdog.com/scrape?api_key=&lt;YOUR_API_KEY&gt;&amp;url=https://data.worldbank.org/indicator/SH.XPD.CHEX.PC.CD?most_recent_value_desc=false&amp;dynamic=true').text</a>

this will provide you with an HTML code of that target URL. Now, you have to use BeautifulSoup to parse HTML.

soup = BeautifulSoup(r,’html.parser’)
country=list()
expense=list()

I have declared two empty lists to store the country names and expenses of each country in 24 hours.

As you can see each country is stored in an “item” tag. We’ll keep all the item tags within a list.

try:
 Countries=soup.find_all(“div”,{“class”:”item”})
except:
 Countries=None

Since there are 190 countries in the world. We are going to run a for loop for each of that countries.

for i in range(0,190):
  country.append(Countries[i+1].find_all(“div”,{“class”:None})[0].text.replace(“\n”,””))
  expense.append(round(float(Countries[i+1].find_all(“div”,{“class”:None})[2].text.replace(“\n”,””).replace(‘,’,’’)))/365)

Data = {‘country’: country,’expense’: expense}

I have divided the expense by 365 because I want to see how these countries spent money on an everyday basis. Obviously, this could have been easier if I have directly divided the given data by 365 but then there is no point in learning right? Now, this is what “Data” looks like.

{'country': ['Central African Republic', 'Burundi', 'Mozambique', 'Congo, Dem. Rep.', 'Gambia, The', 'Niger', 'Madagascar', 'Ethiopia', 'Malawi', 'Mali', 'Eritrea', 'Benin', 'Chad', 'Bangladesh', 'Tanzania', 'Guinea', 'Uganda', 'Haiti', 'Togo', 'Guinea-Bissau', 'Pakistan', 'Burkina Faso', 'Nepal', 'Mauritania', 'Rwanda', 'Senegal', 'Papua New Guinea', 'Lao PDR', 'Tajikistan', 'Zambia', 'Afghanistan', 'Comoros', 'Myanmar', 'India', 'Cameroon', 'Syrian Arab Republic', 'Kenya', 'Ghana', "Cote d'Ivoire", 'Liberia', 'Djibouti', 'Congo, Rep.', 'Yemen, Rep.', 'Kyrgyz Republic', 'Cambodia', 'Nigeria', 'Timor-Leste', 'Lesotho', 'Sierra Leone', 'Bhutan', 'Zimbabwe', 'Angola', 'Sao Tome and Principe', 'Solomon Islands', 'Vanuatu', 'Indonesia', 'Vietnam', 'Philippines', 'Egypt, Arab Rep.', 'Uzbekistan', 'Mongolia', 'Ukraine', 'Sudan', 'Iraq', 'Sri Lanka', 'Cabo Verde', 'Moldova', 'Morocco', 'Fiji', 'Kiribati', 'Nicaragua', 'Guyana', 'Honduras', 'Tonga', 'Bolivia', 'Gabon', 'Eswatini', 'Thailand', 'Jordan', 'Samoa', 'Guatemala', 'St. Vincent and the Grenadines', 'Tunisia', 'Algeria', 'Kazakhstan', 'Azerbaijan', 'Albania', 'Equatorial Guinea', 'El Salvador', 'Jamaica', 'Belize', 'Georgia', 'Libya', 'Peru', 'Belarus', 'Paraguay', 'North Macedonia', 'Colombia', 'Suriname', 'Armenia', 'Malaysia', 'Botswana', 'Micronesia, Fed. Sts.', 'China', 'Namibia', 'Dominican Republic', 'Iran, Islamic Rep.', 'Dominica', 'Turkmenistan', 'South Africa', 'Bosnia and Herzegovina', 'Mexico', 'Turkey', 'Russian Federation', 'Romania', 'St. Lucia', 'Serbia', 'Ecuador', 'Tuvalu', 'Grenada', 'Montenegro', 'Mauritius', 'Seychelles', 'Bulgaria', 'Antigua and Barbuda', 'Brunei Darussalam', 'Oman', 'Lebanon', 'Poland', 'Marshall Islands', 'Latvia', 'Croatia', 'Costa Rica', 'St. Kitts and Nevis', 'Hungary', 'Argentina', 'Cuba', 'Lithuania', 'Nauru', 'Brazil', 'Panama', 'Maldives', 'Trinidad and Tobago', 'Kuwait', 'Bahrain', 'Saudi Arabia', 'Barbados', 'Slovak Republic', 'Estonia', 'Chile', 'Czech Republic', 'United Arab Emirates', 'Uruguay', 'Greece', 'Venezuela, RB', 'Cyprus', 'Palau', 'Portugal', 'Qatar', 'Slovenia', 'Bahamas, The', 'Korea, Rep.', 'Malta', 'Spain', 'Singapore', 'Italy', 'Israel', 'Monaco', 'San Marino', 'New Zealand', 'Andorra', 'United Kingdom', 'Finland', 'Belgium', 'Japan', 'France', 'Canada', 'Austria', 'Germany', 'Netherlands', 'Ireland', 'Australia', 'Iceland', 'Denmark', 'Sweden', 'Luxembourg', 'Norway', 'Switzerland', 'United States', 'World'], 'expense': [0.043835616438356165, 0.049315068493150684, 0.052054794520547946, 0.057534246575342465, 0.057534246575342465, 0.06301369863013699, 0.06575342465753424, 0.07671232876712329, 0.0821917808219178, 0.0821917808219178, 0.0821917808219178, 0.0821917808219178, 0.08767123287671233, 0.09315068493150686, 0.09863013698630137, 0.10136986301369863, 0.10410958904109589, 0.10410958904109589, 0.10684931506849316, 0.10684931506849316, 0.1095890410958904, 0.11232876712328767, 0.1232876712328767, 0.12876712328767123, 0.13150684931506848, 0.14520547945205478, 0.1506849315068493, 0.1506849315068493, 0.15342465753424658, 0.15616438356164383, 0.15616438356164383, 0.16164383561643836, 0.16986301369863013, 0.1726027397260274, 0.17534246575342466, 0.18082191780821918, 0.18082191780821918, 0.1863013698630137, 0.1863013698630137, 0.1863013698630137, 0.1917808219178082, 0.1917808219178082, 0.19726027397260273, 0.2, 0.2136986301369863, 0.21643835616438356, 0.2191780821917808, 0.2356164383561644, 0.2356164383561644, 0.2493150684931507, 0.25753424657534246, 0.2602739726027397, 0.2876712328767123, 0.29041095890410956, 0.3013698630136986, 0.30684931506849317, 0.336986301369863, 0.35342465753424657, 0.3589041095890411, 0.3698630136986301, 0.3863013698630137, 0.3863013698630137, 0.41643835616438357, 0.4191780821917808, 0.4191780821917808, 0.43561643835616437, 0.4684931506849315, 0.4684931506849315, 0.4931506849315068, 0.5150684931506849, 0.5150684931506849, 0.5260273972602739, 0.547945205479452, 0.5561643835616439, 0.5835616438356165, 0.6027397260273972, 0.6054794520547945, 0.6082191780821918, 0.6136986301369863, 0.6219178082191781, 0.6602739726027397, 0.684931506849315, 0.7013698630136986, 0.7123287671232876, 0.7178082191780822, 0.7342465753424657, 0.7452054794520548, 0.7698630136986301, 0.8054794520547945, 0.810958904109589, 0.8328767123287671, 0.8438356164383561, 0.8575342465753425, 0.8657534246575342, 0.8712328767123287, 0.8958904109589041, 0.8986301369863013, 0.9315068493150684, 0.9753424657534246, 0.9835616438356164, 0.9917808219178083, 1.0410958904109588, 1.0602739726027397, 1.0904109589041096, 1.104109589041096, 1.1342465753424658, 1.1369863013698631, 1.1479452054794521, 1.158904109589041, 1.1726027397260275, 1.2164383561643837, 1.2657534246575342, 1.284931506849315, 1.284931506849315, 1.3041095890410959, 1.3424657534246576, 1.3534246575342466, 1.3835616438356164, 1.389041095890411, 1.4136986301369863, 1.4575342465753425, 1.515068493150685, 1.6356164383561644, 1.6767123287671233, 1.7068493150684931, 1.7287671232876711, 1.7753424657534247, 1.8136986301369864, 2.2164383561643834, 2.3315068493150686, 2.3945205479452056, 2.421917808219178, 2.4356164383561643, 2.5506849315068494, 2.5835616438356164, 2.6164383561643834, 2.66027397260274, 2.706849315068493, 2.7726027397260276, 2.7835616438356166, 2.852054794520548, 2.871232876712329, 2.915068493150685, 2.926027397260274, 3.010958904109589, 3.1424657534246574, 3.1890410958904107, 3.23013698630137, 3.2465753424657535, 3.263013698630137, 3.621917808219178, 3.6246575342465754, 3.778082191780822, 4.13972602739726, 4.323287671232877, 4.476712328767123, 4.586301369863014, 4.934246575342466, 5.005479452054795, 5.024657534246575, 5.027397260273973, 5.6, 6.3780821917808215, 6.5479452054794525, 6.745205479452054, 7.504109589041096, 7.772602739726027, 8.054794520547945, 8.254794520547945, 10.26027397260274, 10.506849315068493, 10.843835616438357, 11.27945205479452, 11.367123287671232, 11.597260273972603, 11.67945205479452, 12.213698630136987, 12.843835616438357, 12.915068493150685, 12.991780821917809, 13.038356164383561, 13.704109589041096, 13.873972602739727, 15.24931506849315, 15.646575342465754, 17.18082191780822, 20.487671232876714, 26.947945205479453, 27.041095890410958, 2.8109589041095893]}

Dataframe

Before even starting to plot a graph we have to prepare a DataFrame using pandas. Now, if you don’t know what is DataFrame, DataFrame is a two-dimensional size-mutable, potentially heterogeneous tabular data structure with labeled axes (rows and columns). I know you didn’t get it right. So, just read this article, this will help you a lot. Creating one is very simple and straightforward.

df = pd.DataFrame(Data,columns =[‘country’, ‘expense’])

Visualization

This project is indicative of data science because the majority of time was spent collecting and formatting the data. However, now that we have a clean dataset, we get to make some plots! We can use both matplotlib and seaborn to visualize the data. If we aren’t too concerned about aesthetics, we can use the built-in dataframe plot method to show results quickly:

df.plot(kind = ‘bar’, x =’country’, y=’expense’)
plt.show()

I know the names of the countries are pretty small. But you can download and analyze it. Now, the main thing you can see is that many countries are spending less than a dollar, which is pretty shocking. The majority of Countries are from Asia and Africa. In my opinion, WHO should focus more on these countries rather than developed countries in the West. This is not necessarily a publication-worthy plot, but it’s a nice way to wrap up a small project.

Conclusions

The most effective way to learn technical skills is by doing. While this whole project could have been done manually by inserting values into Excel, I like to take the long view and think about how the skills learned here will help in the future. The process of learning is more important than the final result, and in this project, we were able to see how to use 3 critical skills for data science:

1. Web Scraping: Retrieving online data
2. BeautifulSoup: Parsing our data to extract information
3. Visualization: Showcasing all our hard work

Now, get out there and start your own project and remember: it doesn’t have to be world-changing to be worthwhile. Feel free to comment and ask me anything. You can follow me on Twitter and Medium.

Additional Resources

And there’s the list! At this point, you should feel comfortable writing your first web scraper to gather data from any website. Here are a few additional resources that you may find helpful during your web scraping journey:

• List of web scraping proxy services
• Getting Started with Selenium and Python
• BeautifulSoup Documentation
• Scrapingdog Documentation
• Guide to web scraping