Omdena Chapter Page: Pakistan

Omdena Pakistan Chapter - Omdena Chapters
Completed Project(s)
1st Challenge Completed:

AI Applied: Reducing the Energy Crises in Pakistan with Machine Learning

The Problem 

Pakistan has a population of 224m people yet only half have access to 24/7 electricity. Without electricity, there are no computers or the internet. There are no fridges to keep food fresh. There is no electric water pump. There is nowhere to charge a mobile phone.

Schools and Hospitals struggle to provide basic services. Widening electricity access is an essential first step for improving education, healthcare, and local economies.

Centralized planning for electricity in Pakistan has failed. The government has built-in failure by fixing prices and profits; there is widespread corruption, and banks will not lend to new power plants. Meanwhile, half the existing plants lie idle, and the rest operate below capacity.

Millions live under the grid but not connected to it; previously connected but some equipment failed and has not been replaced, or they have electricity but only enough for a light bulb; or they have it, but it is unreliable due to daily power cuts.

The Project Goals

1. To find target sites we need to exclude those that already have electricity. In addition, those close to the grid were given low priority as they are more likely to receive it directly in the future. The volume of available and free satellite data is incredible. There are night-time light images that clearly show towns that have light.

2. But how do we validate that? Here we can leverage the magic of google maps. I find it awesome to be able to zoom in on a road in Pakistan to see whether it has streetlamps. Based on a selection of test towns it was possible to calibrate and validate the data from satellite images.

3. For the electricity grid, you may think the government and electricity companies would know where their cables are. But they do not!

4. Fortunately, we could leverage an existing model to identify electricity cables that used a combination of machine learning on satellite images and human checking.

The Learning Outcomes

1. Data Extraction

2. Data preprocessing

3. Data Annotation

4. Classification with ML Models

5. Data Visualization

Project Application: https://renew-rho.vercel.app/

Source Code: https://github.com/OmdenaAI/omdena-pakistan-energy-crisis

Link to the Original project: Harnessing AI for Renewable Energy access in Africa

 

 

2nd Challenge Completed:

 

COVID’19 Analysis in Pakistan

The Background

The economic impact of the COVID-19 pandemic in Pakistan has been largely disruptive. According to the Ministry of Statistics, Pakistan’s growth in the fourth quarter of the fiscal year of 2020 went down by 3.1%.

If we could map the shortage and unavailability of vaccines across various states in Pakistan vs the wastage of vaccines across the country, this could help in increasing the efficiency of the vaccination. The current total vaccinated Pakistani across the whole country (shown below), (by our World in Data)

A large proportion of the covid19 mortality rate in Pakistan is because of failing to early identify the presence of coronavirus and its symptoms. Identifying symptoms of Covid19 and mapping its severity level not only help diagnose the viral disease faster and efficiently but also help to reduce the mortality rate by getting medication much sooner.

 

The Problem

In this 4-weeks project, the goal is to build a model to analyse and model the change in economic trends over the past couple of years in Pakistan due to the outbreak of covid and correlate it with various factors that contribute to the rise and fall of the economy. The NLP-based model can then be used to predict the change in future trends in the economy based on these factors. The sentiment of people is modelled from the historical data which can then be used to predict the real-time sentiment and the factors contributing to that sentiment when a potential third wave or a new outbreak happens within the society.

Other objectives include building computer vision models to early detect the presence of covid through X-ray images or CT scans and classify the severity of Covid-19. The project will be made open-source and the results obtained can help create awareness about the spread of Covid19 and its impact on the Pakistan economy. The final tracker and the dashboard made can be used as a tool to track vaccine availability across various states

 

The Project Goals

Economic Impact of Covid19 in Pakistan

1. Collect covid related data from the various up to date sources

2. Data Pre-processing and Exploratory Analysis

3. An interactive Plot /Map displaying current covid analytics of each state.

4. Forecast the relation between covid vs economic change in past time and estimate the effect of covid in the future through Time-series analysis.

Shortage and wastage of Covid19 vaccines 

1. Gather and pre-process the available public open-source data on the latest statistics on Covid19 vaccines.

2. NLP-based Tracker to track the vaccinated population.

3. Interactive Dashboard showing the stepwise-wise statistics of active cases, death cases are given the time range, age-group specific distribution.

4. Correlation matrix between total people vaccinated vs total vaccine produced.

5. Correlation matrix between coronavirus (Covid-19), mortality and morbidity rate, burden over radiologists.

6. A Knowledge Graph-based solution to visualise the state-wide shortage and availability of COVID19 vaccine.

Early prediction of Covid19 Through Computer Vision

1. Predict the likelihood of covid positivity from chest CT Scan images and respiratory patterns. From demographic and clinical data (the patient’s age and sex, exposure history, symptoms, and laboratory tests) were put into a machine-learning model to classify COVID-19 positivity.

2. Using model predictions to track patient recovery from Multiple CT scans taken during treatment can be used to analyse whether the patient is recovering or worsening.

3. A web application to visualize the severity of covid from chest X-ray images.

 

The Tasks & Timeline

Week 1 Week 2 Week 3 Week 4

– NLP text Pre-processing

-Twitter Hashtag extraction

-Image pre-processing

– Exploratory Data Analysis (EDA)

-Interactive plots with Real-time data

-Interactive map vaccination tracker 

-Classification model based on covid symptoms

-Sentiment Analysis Model 

-Image Classification model based on symptoms

-Start building Streamlit WebApp

-Finish Integrating WebApp

-Deploy the App in Cloud Application Platforms

 

Learning Sessions and Timeline:

1. EDA Workshops: Starts from Sept 25th

2. Streamlit Bootcamp: Starts from 11th Oct

The Learning Outcomes

1. Data Pre-processing

2. NLP based EDA

3. Developing and Deploying Dashboards

4. Time Series Forecasting

5. Computer vision and model Building

 

 

Source code 

https://github.com/OmdenaAI/omdena-pakistan-covid-analysis


Dashboard

https://public.tableau.com/app/profile/nadia2778/viz/covid19-pakistan/Dashboard2?publi[…]AR3nROlMTYIzncLMnGfrKV3ptvL_Ki6B4zTyB5L8u1Oj-tGRWL2gf7Ua14k by “Nadia Nizam”

https://share.streamlit.io/muhammadawon/covid-radiography-classifier/main/covid-webapp/app.py by “Muhammad Awon”

https://share.streamlit.io/usmanes70/streamlit-app/main/pakistan.py by “Usman Ayaz”

We will be running an AI project soon…. Stay Tuned!

Pakistan Chapter
 
Pakistan Chapter Leads

Qasim Hassan

Qasim Hassan is a student/entrepreneur visionary passionate individual with a strategic mindset toward  Artificial Intelligence & Data Science. He is a Jr. Machine Learning Engineer at Omdena and formally a Jr. Data Scientist at AISOL.

Our Partner(s):