Pandas Dataframe to JSON

 Pandas has a function (.to_json()) that takes a series/dataframe and converts it into JSON object.

It has an interesting parameter (orient) that allow different orientations of the JSON string format. Allowed values are: {'split','records','index','columns','values','table'}. Let see how it works.

Assuming we have the table below in out dataframe and now we want it as a JSON data for use on another web platform (JSON work better on web than tabular data).

As usual, we read the data table into a dataframe variable (here it is called: json_df).

Then the .to_json() function is called on the dataframe to convert it to JSON like this:-

json_df.to_json()

There are many parameters we can pass into the function such as path to save the JSON and orient to orient the JSON in different formats. Below, we will explore the various orient options available.

As you can see, each orient value has its own different JSON format or structure.

print(json_df.to_json(orient='columns'))
output = {"Question":{"0":"In what country is the city 'Tokyo'?","1":"In what country is the city 'Delhi'?","2":"In what country is the city 'Cairo'?","3":"In what country is the city 'Chongqing'?","4":"In what country is the city 'Orlando'?","5":"In what country is the city 'Abuja'?"},"Option 1":{"0":"Japan","1":"India","2":"Argentina","3":"Philippines","4":"Mexico","5":"Spain"},"Option 2":{"0":"India","1":"United States","2":"Bangladesh","3":"Pakistan","4":"Angola","5":"Nigeria"},"Option 3":{"0":"Brazil","1":"China","2":"Egypt","3":"China","4":"United States","5":"Canada"}}


print(json_df.to_json(orient='split'))
output = {"columns":["Question","Option 1","Option 2","Option 3"],"index":[0,1,2,3,4,5],"data":[["In what country is the city 'Tokyo'?","Japan","India","Brazil"],["In what country is the city 'Delhi'?","India","United States","China"],["In what country is the city 'Cairo'?","Argentina","Bangladesh","Egypt"],["In what country is the city 'Chongqing'?","Philippines","Pakistan","China"],["In what country is the city 'Orlando'?","Mexico","Angola","United States"],["In what country is the city 'Abuja'?","Spain","Nigeria","Canada"]]}


print(json_df.to_json(orient='records'))
output = [{"Question":"In what country is the city 'Tokyo'?","Option 1":"Japan","Option 2":"India","Option 3":"Brazil"},{"Question":"In what country is the city 'Delhi'?","Option 1":"India","Option 2":"United States","Option 3":"China"},{"Question":"In what country is the city 'Cairo'?","Option 1":"Argentina","Option 2":"Bangladesh","Option 3":"Egypt"},{"Question":"In what country is the city 'Chongqing'?","Option 1":"Philippines","Option 2":"Pakistan","Option 3":"China"},{"Question":"In what country is the city 'Orlando'?","Option 1":"Mexico","Option 2":"Angola","Option 3":"United States"},{"Question":"In what country is the city 'Abuja'?","Option 1":"Spain","Option 2":"Nigeria","Option 3":"Canada"}]


print(json_df.to_json(orient='values'))
output = [["In what country is the city 'Tokyo'?","Japan","India","Brazil"],["In what country is the city 'Delhi'?","India","United States","China"],["In what country is the city 'Cairo'?","Argentina","Bangladesh","Egypt"],["In what country is the city 'Chongqing'?","Philippines","Pakistan","China"],["In what country is the city 'Orlando'?","Mexico","Angola","United States"],["In what country is the city 'Abuja'?","Spain","Nigeria","Canada"]]


print(json_df.to_json(orient='table'))
output = {"schema": {"fields":[{"name":"index","type":"integer"},{"name":"Question","type":"string"},{"name":"Option 1","type":"string"},{"name":"Option 2","type":"string"},{"name":"Option 3","type":"string"}],"primaryKey":["index"],"pandas_version":"0.20.0"}, "data": [{"index":0,"Question":"In what country is the city 'Tokyo'?","Option 1":"Japan","Option 2":"India","Option 3":"Brazil"},{"index":1,"Question":"In what country is the city 'Delhi'?","Option 1":"India","Option 2":"United States","Option 3":"China"},{"index":2,"Question":"In what country is the city 'Cairo'?","Option 1":"Argentina","Option 2":"Bangladesh","Option 3":"Egypt"},{"index":3,"Question":"In what country is the city 'Chongqing'?","Option 1":"Philippines","Option 2":"Pakistan","Option 3":"China"},{"index":4,"Question":"In what country is the city 'Orlando'?","Option 1":"Mexico","Option 2":"Angola","Option 3":"United States"},{"index":5,"Question":"In what country is the city 'Abuja'?","Option 1":"Spain","Option 2":"Nigeria","Option 3":"Canada"}]}


print(json_df.to_json(orient='index'))
output = {"0":{"Question":"In what country is the city 'Tokyo'?","Option 1":"Japan","Option 2":"India","Option 3":"Brazil"},"1":{"Question":"In what country is the city 'Delhi'?","Option 1":"India","Option 2":"United States","Option 3":"China"},"2":{"Question":"In what country is the city 'Cairo'?","Option 1":"Argentina","Option 2":"Bangladesh","Option 3":"Egypt"},"3":{"Question":"In what country is the city 'Chongqing'?","Option 1":"Philippines","Option 2":"Pakistan","Option 3":"China"},"4":{"Question":"In what country is the city 'Orlando'?","Option 1":"Mexico","Option 2":"Angola","Option 3":"United States"},"5":{"Question":"In what country is the city 'Abuja'?","Option 1":"Spain","Option 2":"Nigeria","Option 3":"Canada"}}

The structure you will adopt will depend on where you want to use it. Usually, you would find sample JSON structure from the platform you intend to use. Compare the option and adopt the one that suit your project.

That is it!

Pandas dataframe.append() function - Append 'Jeo Biden' to US presidents dataframe

Here we got a dataframe of 45 USA presidents with four columns as seen below.

The dataframe is this wikipedia table that lists the presidents of the United States. As you already saw, the wikipedia table  has been updated. So, we need to update our dataframe with 46th president.

Now we got a new president-elect (Joseph Robinette Biden Jr.), we will create a new dataframe for him and use the append() function to add it to the end of our dataframe.

To create the Biden's dataframe, we store his data in a dictionary and use it to make a one row dataframe to be appended to the main dataframe above.

biden_data = {
            'Name':['Joseph Robinette Biden Jr.'], 
              'Image':['https://upload.wikimedia.org/wikipedia/commons/9/99/Joe_Biden_official_portrait_2013_cropped_%28cropped%29.jpg'], 
              'Party':['Democratic'], 
              'Presidency Period':['20-Jan-21 – Incumbent']
             }

biden_df = pd.DataFrame(biden_data)
biden_df

Call the append function on the USA presidents dataframe like this:- You'll notice that the index is '0' for the total line. We want to change that from '0' to '46' using rename.

us_presidents_df.append(biden_df).rename(index={0:"45"})

That is it!

Mass make folders from list of states

 I have written on this before at How to make Multiple Folders/Directories in Python. Refer to that page for more detailed instructions.

Anyway, here I have a list of states in Nigeria that I want to generate folder for each.

import os
states = ['Ebonyi', 'Edo', 'Ekiti', 'Enugu', 'Fct', 'Gombe', 'Imo', 'Jigawa', 'Kaduna', 'Katsina', 'Kebbi', 'Kogi', 'Kwara', 'Lagos', 'Nasarawa', 'Niger', 'Ogun', 'Ondo', 'Osun', 'Akwa Ibom', 'Oyo', 'Plateau', 'Rivers', 'Sokoto', 'Taraba', 'Yobe', 'Zamfara', 'Anambra', 'Bauchi', 'Bayelsa', 'Benue', 'Borno', 'Cross River', 'Delta']

for n in states: # Looping over the individual state to make folder name
    if type(n) == str:
        print ('Making directory for.... ', n)

        directory = n + '__state'
        if not os.path.exists(directory):
            os.makedirs(directory)

The resulting output is seen below:-

Enjoy!

Format number to thousands string ₦ currency using Python f-string

 Here we have list of digits representing prices of items in Nigeria Naira (₦) currency like this:-

money_naira = [2495, 93988, 39118, 19973, 39579, 35723, 80216, 56725, 16132, 82275, 18439, 34919, 17117, 85879, 51153, 7737, 35367, 9753, 86648, 87650, 58011, 2219, 1768, 8612, 2901, 5041, 3405, 8486, 7742, 5008, 7150, 5553, 9320, 2736, 9151, 9894, 2812, 6466, 1194, 4322, 6696, 6144, 6227, 2479, 3027, 4052, 7580, 1736, 9979, 1638, 2369, 8702, 1353, 9695, 4072, 4065, 7742, 7887, 7620]

But there is a problem! These number doesn't reads well by humans, it is just a numeric digit that can mean anything. So, we have to represent it in such a way that readers can tell they are actual money in Nigeria Naira (₦).

That is to say; 2495 will be presented as ₦2,495.00

This can be achieved using python string formatting (f-string or format()). f-string is the newest way to format string in python, it is available since Python 3.6. You read more from this f-string tutorial.

Lets see how it done.

for money in money_naira:
    # Using f-string
    print(f"₦{money:,.2f}")
    
    # Using format() function
    print("₦{:,.2f}".format(money))

The output is:

₦2,495.00, ₦93,988.00, ₦39,118.00, ₦19,973.00, ₦39,579.00, ₦35,723.00, ₦80,216.00, ₦56,725.00, ₦16,132.00, ₦82,275.00, ₦18,439.00, ₦34,919.00, ₦17,117.00, ₦85,879.00, ₦51,153.00, ₦7,737.00, ₦35,367.00, ₦9,753.00, ₦86,648.00, ₦87,650.00, ₦58,011.00, ₦2,219.00, ₦1,768.00, ₦8,612.00, ₦2,901.00, ₦5,041.00, ₦3,405.00, ₦8,486.00, ₦7,742.00, ₦5,008.00, ₦7,150.00, ₦5,553.00, ₦9,320.00, ₦2,736.00, ₦9,151.00, ₦9,894.00, ₦2,812.00, ₦6,466.00, ₦1,194.00, ₦4,322.00, ₦6,696.00, ₦6,144.00, ₦6,227.00, ₦2,479.00, ₦3,027.00, ₦4,052.00, ₦7,580.00, ₦1,736.00, ₦9,979.00, ₦1,638.00, ₦2,369.00, ₦8,702.00, ₦1,353.00, ₦9,695.00, ₦4,072.00, ₦4,065.00, ₦7,742.00, ₦7,887.00, ₦7,620.00

₦2,495.00, ₦93,988.00, ₦39,118.00, ₦19,973.00, ₦39,579.00, ₦35,723.00, ₦80,216.00, ₦56,725.00, ₦16,132.00, ₦82,275.00, ₦18,439.00, ₦34,919.00, ₦17,117.00, ₦85,879.00, ₦51,153.00, ₦7,737.00, ₦35,367.00, ₦9,753.00, ₦86,648.00, ₦87,650.00, ₦58,011.00, ₦2,219.00, ₦1,768.00, ₦8,612.00, ₦2,901.00, ₦5,041.00, ₦3,405.00, ₦8,486.00, ₦7,742.00, ₦5,008.00, ₦7,150.00, ₦5,553.00, ₦9,320.00, ₦2,736.00, ₦9,151.00, ₦9,894.00, ₦2,812.00, ₦6,466.00, ₦1,194.00, ₦4,322.00, ₦6,696.00, ₦6,144.00, ₦6,227.00, ₦2,479.00, ₦3,027.00, ₦4,052.00, ₦7,580.00, ₦1,736.00, ₦9,979.00, ₦1,638.00, ₦2,369.00, ₦8,702.00, ₦1,353.00, ₦9,695.00, ₦4,072.00, ₦4,065.00, ₦7,742.00, ₦7,887.00, ₦7,620.00

That is it!

Javascript array Vs Python list

Introduction

In this article, I will talk about one of the most important and similar data structure in both Javascript and Python.

Javascript array (similar to Python list) is an object that lets you store multiple values in a single variable. For example, an array/list of countries would look like this: ['Nigeria', 'Canada', 'England', 'Mexico', 'China', 'India', 'Kuwait']. While this array/list is made up of only string data type, it can contain multiple data types.

Lets take a look at some commonly used methods when working with array/list in both Javascript and Python.

Creating array/list

There are several ways of creating an array/list in both JS and Py. We will see some of the common ways here.

JS:

let countries0 = ['Nigeria', 'Canada', 'England', 'Mexico', 'China', 'India', 'Kuwait'];

let countries1 = new Array('Nigeria', 'Canada', 'England', 'Mexico', 'China', 'India', 'Kuwait'); 

let countries2 = new Array()
countries1[0] = 'Nigeria';
countries1[1] = 'Canada';
countries1[3] = 'England';

Py:

countries = ['Nigeria', 'Canada', 'England', 'Mexico', 'China', 'India', 'Kuwait']

The list() function in python is used to convert from other types to list.

Screenshots of Google maps at different zoom levels

 The scenario here is that we have latitude and longitude coordinates of cities that we want their Google maps screenshots at different zoom levels.

As at the time of writing, the construct for google maps url is like this: https://www.google.ng/maps/@latitude,longitude,zoomz >>>> https://www.google.ng/maps/@9.057809,7.4903376,15z

So, at that city coordinate we want to take the screenshots at varying zooms from 6 to 20. The url construct will be:-

• https://www.google.ng/maps/@9.057809,7.4903376,6z
• https://www.google.ng/maps/@9.057809,7.4903376,7z
• https://www.google.ng/maps/@9.057809,7.4903376,8z
• https://www.google.ng/maps/@9.057809,7.4903376,9z
• https://www.google.ng/maps/@9.057809,7.4903376,10z
• https://www.google.ng/maps/@9.057809,7.4903376,11z
• https://www.google.ng/maps/@9.057809,7.4903376,12z
• https://www.google.ng/maps/@9.057809,7.4903376,13z
• https://www.google.ng/maps/@9.057809,7.4903376,14z
• https://www.google.ng/maps/@9.057809,7.4903376,15z
• https://www.google.ng/maps/@9.057809,7.4903376,16z
• https://www.google.ng/maps/@9.057809,7.4903376,17z
• https://www.google.ng/maps/@9.057809,7.4903376,18z
• https://www.google.ng/maps/@9.057809,7.4903376,19z
• https://www.google.ng/maps/@9.057809,7.4903376,20z

As you would have noticed from the url above, only the zoom level changes. The city coordinate remains the same for the fifteen zoom levels (6z - 20z).

Lets write a python script that will handle this for us.

1) Generate the URLs

base_url = 'https://www.google.ng/maps/@9.057809,7.4903376,{}z'

for x in range(6, 21):
    print(base_url.format(str(x)))

2) Take screenshots

Here I used selenium module to open and take screenshot of each map url. I also used the time module to make two second delays to allow the map load completely before taking the screenshots.

Also note that I saved the urls in a list called: url_list

import time
from selenium import webdriver


# Load chrome browser driver
chrome_driver = 'C:\\Users\\Yusuf_08039508010\\Documents\\chromedriver.exe'
driver = webdriver.Chrome(chrome_driver)


for url in url_list:
    driver.get(url)
    time.sleep(2)
    
    img_name = url.split(',')[-1]
    driver.save_screenshot(img_name + ".png")
    
    print('Saving image...', img_name)

That is it!

Split dataframe into chunks

The dataset we will use for this demo is the sample 'World Cities Database' from Simplemaps.com. Download a copy from the given url above.

Upon inspecting the file, you will see that there are 15493 rows and 11 columns. 

There several ways to split a dataframe into parts or chunks. How it is split depends on how to dataframe want to be used. In this post we will take a look at three ways a dataframe can be split into parts. Specifically, we will look at:-

Mp3 of Nigeria state names using python Text to Speech (TTS) library

I was working on this project, where I needed an audio file to say something like "Kaduna state has a population of 6,113,503".

Doing this manually by recording the mp3 and saving it for the whole 30+ states is not going to be an easy task, so I decided to look for a script solution. Luckily, I found a python library that can do it. 

There are several TTS libraries in python. The one am going to use is called "pyttsx3" which works without internet connection and supports multiple TTS engines including Sapi5, nsss, and espeak.

Lets get our hands dirty.

I made use of this table (List of Nigerian states by population) on wikipedia. To get the table data, lets use pandas read_html() function which return list of table from HTML as dataframes. The table we are interested in is the first table, so it is the first dataframe.

import pandas as pd

# Read table from HTML....
df = pd.read_html('https://en.wikipedia.org/wiki/List_of_Nigerian_states_by_population')
df[0]

import pandas as pd
import pyttsx3


# Note that df is a list of dataframe....

# Construct the TTS string from the df[0] that say: "<XXX State> has a population of <Population>"
# df[0]['State'][0] + ' has a population of ' + format(df[0]['Population (2006)'][0], ',')

# Using for loop...
tts_list = []
i, j = 0, 0
for item in df[0]['State']:
    tts = df[0]['State'][i] + ' has a population of ' + format(df[0]['Population (2006)'][j], ',')
    tts_list.append(tts)
    
    i += 1
    j += 1
    

# Using list comprehension with zip()...
tts_list = [ i + ' has a population of ' + format(j, ',') for i, j in zip(df[0]['State'], df[0]['Population (2006)']) ]


# We now have a list of all the string we wanted, so lets use the list to create mp3...
engine = pyttsx3.init()

i = 0
for s in tts_list:
    # engine.say(s) # Speak the mp3
    engine.save_to_file(s, df[0]['State'][i]+'.mp3') # say mp3 to file/folder
    engine.runAndWait()
    
    i += 1

The pyttsx3 module can do a lot more including setting volume and setting voices for male/female etc.

You can download the MP3 files here.

Enjoy!

QGIS Attribute table Conditional formatting

Not in all instances that we want to run query to select items on attribute table. Sometimes we just want to highlight of format the attribute table so that certain item that met a specified characteristics are noticeable on the table.

For example, we may want to see rows that have duplicate records or rows whose name ends with a certain alphabet. Whatever the case maybe, QGIS has a handy tool called "Conditional formatting" that allow you format your attribute table as you wanted. Let see some few examples:-

Conditional formatting base on duplicate records

We can use the count() function which returns the count of matching features to check for duplicte records. To do this, we check if the count is greater than one, if so than the record appeared more than once and hence it is duplicated.

The expression is: count("Name" , "Name") > 1 where "Name" is the attribute field name. We can select the duplicated attributes as seen below. However, what we wanted here is to conditionally format the table base on duplicate records.

To achieve conditional formatting, click on the "conditional formatting" icon as seen below. Then add a new rule. Note, you can format just the field or the full row.

Add a name, enter the conditional expression (in this case it is: count("Name" , "Name") > 1). Then set the background and text colors. After which you will click on 'Done' button to apply the formatting.

Conditional formatting base on name ending with 'a'

Here we will use the regexp_match() function to find names that end with alphabet "a". The correct expression is: regexp_match("Name", 'a$'), this has to do with regular expressions.

We will add this formatting to the one above. So, we just need to add another "New Rule", to format items under 'Name' field/column that ends with 'a'.

We now have colors separating all the criteria we wanted in a nice color table.

That is it!

Download image files using python requests module

 Here I have list of data with links to their corresponding images. Specifically, this is a list of presidents in the USA and Nigeria.

I have to download the images/photos. So, I wrote a python script that uses the requests library to download the images as seen below:-

There are two approaches when saving an image file (or any file) on the hard disk. One is using the open with close methods or using the with statement.

import requests
import pandas as pd

image_url = "https://upload.wikimedia.org/wikipedia/commons/thumb/7/70/John_Adams%2C_Gilbert_Stuart%2C_c1800_1815.jpg/842px-John_Adams%2C_Gilbert_Stuart%2C_c1800_1815.jpg"

# -------- 1st approach ------------

response = requests.get(image_url)

img_file = open("downloaded_img/sample_image.png", "wb")
img_file.write(response.content)
img_file.close()

# -------- 2nd approach ------------

img_file = requests.get(image_url).content
with open('downloaded_img/image_name.jpg', 'wb') as f:
    f.write(img_file)

Downloading the US presidents photos:-

us_presidents_df = pd.read_csv(r"C:\Users\Yusuf_08039508010\Desktop\US Presidents\USA List.csv")

# Bulk processing....
for name, img in zip(us_presidents_df['Name'], us_presidents_df['Image']):
    print('Processing....', name, img)
    
    img_file = requests.get(img).content
    with open('downloaded_img/US Presidents/'+name+'.jpg', 'wb') as f:
        f.write(img_file)

Downloading Nigerian presidents photos:-

nigeria_presidents_df = pd.read_csv(r"C:\Users\Yusuf_08039508010\Desktop\NG List.csv")

for name, img in zip(nigeria_presidents_df['Name'], nigeria_presidents_df['Image']):
    print('Processing....', name, img)
    
    response = requests.get(img)

    img_file = open('downloaded_img/NG Presidents/'+name+'.png', "wb")
    img_file.write(response.content)
    img_file.close()
    
print('Completed....')

That is it!