Welcome to the Kenya Chapters!
There are 4 active chapters in Kenya:
- Kisumu, Kenya
- Nairobi, Kenya
- Kiambu, Kenya
- Nakuru, Kenya
Apply here to be a chapter lead for other cities and/or universities in Kenya
Upcoming Projects
Nairobi, Kenya Chapter
Project Starts: July 13th
Nairobi, Kenya Local Chapter – Plant rust diseases classification
Background
Presently in Kenya, there are rarely any farmer workshops to educate farmers on the current farming trends and diseases, this reality adversely hinders farming and cultivation of healthy crops to ensure maximum profits. It is our firm belief that the introduction of this technology would promote the introduction and adoption of best farming practices and maximise farmer profits. Such a system would help keep track of crop diseases and best treatment options.
The Problem
Kenya experiences long periods of drought and short periods of rain, this has led to low crop produce and inflation of food prices. Most farmers are not conversant with the best farming practices and are still using old methods. Farmers at times confuse rust diseases with drought symptoms and thus fail to take action and spray the crops or provide the required nutrients.
The Project Goals
Identification Services with Machine Learning
- To develop a recognition model that will identify rust diseases with a high degree of accuracy.
- To Test our Models accuracy and attempt to improve on our model.
- To integrate our final model into a suitable Database, in an application.
- To deploy an API or demo of the proposed system.
- To Test the final Product. Measure its effectiveness.
The Learning Outcomes
- Data collection
- Data Processing
- Labelling of Data
- ML Model for extraction of rust diseases.
- ML Model for identification and comparison of rust diseases against Known Database.
- A Database system for storing the Face-Recognition Database Content.
- Testing of Results and Fine Tuning the model.
- Deployment of the whole system
The Tasks & Timeline
Week 1 | Week 2 | Week 3 | Week 4 | Week 5 | |
1.Data Gathering 2.Understanding the problem | 1.Data Cleaning 2.Pre-processing and analysis. | 1.Understanding the Model. 2. Deeper Into the ML models | 1.Implementing the Model. 2.Fine-Tuning The Model. | 1.Building the Database. 2.Integrating the Database. 3.Deploying the Model. |
Kiambu, Kenya Chapter
Project Starts: June 30th
Kiambu, Kenya Chapter – Building AI-Powered Chatbot to Offer 24/7 Support Services For University Website
Background
Universities in the country are made up of millennial individuals who spend most of their time on the internet browsing and watching fun activities on social media sites. Students always need to spend less time seeking what they want or inquiring about any issues from the internet. Further, the young individuals need interactive sites where they can get information instantly hence the spread of social media sites like WhatsApp, Twitter, and Facebook.
Chatbot technology is a new way of increasing online presence among institutions and reducing time that students spend on the school website seeking information or clarification. Since most students spend evening and night hours visiting different websites for information, such times they are always tired. The need to introduce chatbot that can provide real time responses in an interactive manner would increase the website users seeking different information.
Our institution (Dedan Kimathi University of Technology) is not excluded since it has not adopted the use of chatbot technology on the website to improve student experience. Students are forced to browse the website seeking information that is relevant to them.
Problem Statement
The institution website has a lot of information that is always updated daily for different audiences such as students, staff, and community. Students take time navigating on the school website seeking information that is relevant to them, a norm that could be a challenge for millennial students since they always need information within the shortest time. University students spend most of their evening hours visiting different websites including school websites to seek information regarding their courses of study, this is a time when they are tired after a long day of study.
It will be of great importance to utilise the fast-growing chatbot technology to create an interactive chatbot system that can provide responses to students whenever they want to seek information about in their courses on the website.
The Project Goals
- Collect data that students seek clarification from the school website in JSON Format.
- Prepare the collected information through Natural Language Preprocessing.
- Create a Deep Learning Model that will help provide real time responses from the inquiries of students.
- Create a web dashboard that will be used for testing purposes before real applications.
Deployment on the school website.
The Learning Outcomes
- Data Collection and Cleaning.
- Natural Language Processing Techniques.
- Deep Learning Models
- Web Development
The Tasks & Timeline
Week 1 Data Collection | Week 2 EDA | Week 3 Regression | Week 4 Deployment |
–Data Collection – Data Cleaning | – Data Preprocessing – NLP Techniques | -Developing Deep Learning Model. | -Creating Website Dashboard Week 5
|
Completed Projects
Kisumu, Kenya Challenge: Improving Fishing Activities in Kenya’s Lakes Region using Geo-Data.
Background
Fishing is a major part of the livelihood of the people living in the Kisumu region in Kenya, mainly known as the lake region. Most individuals solely depend on fishing as a source of income making it the largest economic activity in the area.
Problem Statement
Over the years, changes in weather and climate have interrupted and disadvantaged this activity leading to increased poverty in the region. This project aims to look into changes in temperature, rainfall, and wind and their effects on Fishing Activities and coming up with coping and adaptive strategies for the fishers. Climate change is not the only problem faced in the region but also due to increased population in the region over the past years, there have been increased cases of overfishing to meet the increase in population.
The Project Goals
With this project we hope to:
1. Detect and predict the effect of weather and climate on fishing activities.
Understanding the relationship between climate and fishing patterns will go a long way toward coming up with coping and adaptive strategies for the fishers. The prediction will help prepare.
1. Teach members of the community how to access open-source satellite image data, process it, and use it in other applications
2. Building of AI models that work with such data
3. Get members to feel how a project is conceptualized until it’s implemented
The Learning Outcomes
1. Create awareness on open-source data that can be used to solve problems in our community
2. Have a tool that can improve the livelihood of the people of Kisumu who depend on fishing as one of their economic activities
The Tasks & Timeline
Week 1 Data Collection | Week 2 EDA | Week 3 Regression | Week 4 Deployment |
–Weather Data Collection -Satellite image data collection. -Fishing data collection. | – Exploratory Data Analysis(EDA) -Image preprocessing
| -Image Classification -Machine Learning model Creation | -Creation of a web app -Deploy the App in Cloud Application Platforms |
Kiambu, Kenya Chapter - Applying Machine Learning to Understand Different Factors that Influence School Performance
The Background
According to the Kenya National Bureau of Statistics (KNBS) report 2019, Kiambu County has a population of 2,417,735. Kiambu County’s website states that there are 1225 primary and 303 secondary schools. The number of pre-schools has not been declared on their website. The gross and net enrolment rates for primary are 109.6% and 99.7%, while secondary levels are 69.3% and 61.8%, respectively. It is important to note that the increasing population of Kiambu has necessitated more investment in the education sector.
According to Kiambu County’s website, the pre-schools need more investments in public ECD centers to ensure children from poor backgrounds get access to early education without much strain. Primary schools need to invest in providing additional education facilities because of the increasing number of schools going population. On the other hand, secondary schools need significant investment in the education sector to ensure the completion rate reaches 100 percent since it currently stands at 92.5 percent.
Based on the statistics and recommendations outlined coupled with the introduction of the Competency-Based Curriculum (CBC), there is a clear indication that the quality of education in Kiambu needs to be improved. Therefore, this study seeks to investigate the factors and additional resources required to improve the quality of education.
The Problem
With the enrollment of free primary and secondary school education, the number of students enrolled per category has risen. The government has also introduced the CBC, which requires more facilities in schools. The resources in schools are not enough to guarantee better performance. Therefore, this project aims to determine how different factors have affected learning institutions and performance over the years and recommend steps to decide on how to distribute them. The Country claims more resources are needed in schools to improve performance even as the government tries to achieve a 100% transition
The Project Goals
1. Develop a dashboard that shows the facilities to students ratio and the distribution rate of facilities in the three categories of schools in the county.
2. Build a machine learning model that analyzes the performance of the school and the level of facilities in different academic years.
3. Compare performance, number, and type of facilities in public and private schools.
4. Outline recommendations on which facilities the county and other education stakeholders should prioritize to improve school performance and achieve 100% transition across the levels.
The Learning Outcomes
1. Recommend a check sheet that education stakeholders can use while allocating money for infrastructure development in schools.
2. Visualizations of the distribution of facilities, student population, and performance in the county.
The Tasks & Timeline
| Week 1 | Week 2 | Week 3 | Week 4 |
– Data Extraction -Data Understanding –Data Exploration | – Exploratory Data Analysis (EDA) -Interactive plots with Real-time data -Interactive map for the sub-counties – | -Feature Engineering – Model selection techniques Modeling – | – Model Evaluation Report writing – Presentation |
Nairobi, Kenya Challenge: Build Web Application Dashboard To Report The Covid Situation In Kenya
Background
Dashboard and other visualizations are great when communicating insights. They are easier to understand – especially when going through large amounts of data. In this project, our aim is to scrap PDF reports of the updates of the covid situation in Kenya from the Ministry of health website and extract information that we shall store and also present it in the form of a dashboard. We will then automate the process so that we have an updated dashboard each time that can be accessed by anyone.
Problem Statement
In Kenya we get an update everyday regarding the covid situation in the country by the Ministry of Health. The updates contain statistical numbers on parameters such as the number of people vaccinated,new cases,cases per county etc. This information is given as PDF files on the ministry website and going through these files might be cumbersome. Furthermore, comparing progress or milestones can be a challenge.
It is therefore our aim to have this information in the form of a dashboard as it will be easier to interpret and could provide insights to stakeholders on whether we are winning the war on the virus. It could also allow normal citizens in the country to easily understand the covid situation in the country. We shall also make it possible for anyone to download the data and use it for research purposes such as data modeling.
The Project Goals
With this project we hope to:
1. Scrape PDF data from a website
2. Extract data From the obtained PDFs and transform it in a way that it is easier to use
3. Build a dashboard that reports on the covid situation in the country
4. Automate the above process
5. Deploy the web application with real time updates
6. Allow users to be able to download the above data easily
The Learning Outcomes
1. Learn how to scrape data from websites
2. Building insightful Data Visualizations
3. Working with Databases
4. Build end to end scraping data pipelines
5. Build REST API with Flask
6. Deploying a Web Application
The Tasks & Timeline
| Week 1 | Week 2 | Week 3 | Week 4 |
–Understanding of Problem Statement -Downloading of Data -Workshop on how to collaborate using git | – Exploratory Data Analysis(EDA) -Building Dashboard & Visualizations -Introduction to REST APIs & Web Applications | -Building Backend API -Building Front-end -Workshop on Docker And Docker Compose | -Automating the pipeline -Finish Integrating Web App -Deploy the App in Cloud Application Platforms |
Nakuru, Kenya Challenge : Improving Food Security and Crop Yield in Kenya Through Machine Learning
The Background:
Based on data collected during the 2020 short rains assessment, the Kenya Food Security Steering Group (KFSSG) estimates that around 1.4 million Kenyans in arid and semi-arid areas are facing Crisis (IPC Phase 3) or worse outcomes, an increase of 93 percent compared to the preceding long rains season. Cumulatively below-average rainfall across eastern Kenya resulted in a poor harvest in marginal agricultural livelihood zones and declines in rangeland resources in pastoral areas driving Stressed (IPC Phase 2) and Crisis (IPC Phase 3) outcomes across northern and eastern Kenya.
The Problem:
More than 1.4 million Kenyans in arid and semi-arid areas are facing food crises or worse outcomes. Covid-19 control measures, Desert locust invasion, and climate change have negatively impacted crop production and rangeland resource regeneration.
With the help of Machine Learning, farmers should be able to predict weather patterns and conditions in different places in Kenya for the next farming season while promoting a data-driven agricultural system. Data such as soil PH, temperature, and moisture levels, land usage, combined with other data sources from World Bank’s data portal and Kenya Meteorological Department could be processed to show exactly when and where farmers should improvise their farming method, and to know the best crop type to be cultivated. The data will also help to decide where to invest, and make use of unutilized land for farming
The Project Goals
1. Identify unutilized farming land through satellite imaging.
2. Applying Kenyan-based open-source satellite imagery dataset to make crop yield prediction.
3. Create a weather information sharing system for farmers for better farming decisions.
The Tasks & Timeline:
Week 1 | Week 2 | Week 3 | Week 4 |
–Weather Data Collection -Satellite image data collection. | – Exploratory Data Analysis(EDA) -Image preprocessing
| -Image Classification -Machine Learning model Creation | -Creation of a web app -Deploy the App in Cloud Application Platforms |
The Learning Outcomes
1. Satellite Image data collection
2. Weather patterns analysis
3. Computer Vision for crop type detection
4. Data visualization using pandas, matplotlib and QGSI
Kiambu, Kenya Challenge: Applying Machine Learning to understand different factors that influence school performance
– Data Extraction
-Data Understanding
–Data ExplorationBackground
According to the Kenya National Bureau of Statistics (KNBS) report 2019, Kiambu County has a population of 2,417,735. Kiambu County’s website states that there are 1225 primary and 303 secondary schools. The number of pre-schools has not been declared on their website. The gross and net enrolment rates for primary are 109.6% and 99.7%, while secondary levels are 69.3% and 61.8%, respectively. It is important to note that the increasing population of Kiambu has necessitated more investment in the education sector.
According to Kiambu County’s website, the pre-schools need more investments in public ECD centers to ensure children from poor backgrounds get access to early education without much strain. Primary schools need to invest in providing additional education facilities because of the increasing number of schools going population. On the other hand, secondary schools need significant investment in the education sector to ensure the completion rate reaches 100 percent since it currently stands at 92.5 percent.
Based on the statistics and recommendations outlined coupled with the introduction of the Competency-Based Curriculum (CBC), there is a clear indication that the quality of education in Kiambu needs to be improved. Therefore, this study seeks to investigate the factors and additional resources required to improve the quality of education.
Problem Statement
With the enrollment of free primary and secondary school education, the number of students enrolled per category has risen. The government has also introduced the CBC, which requires more facilities in schools. The resources in schools are not enough to guarantee better performance. Therefore, this project aims to determine how different factors have affected learning institutions and performance over the years and recommend steps to decide on how to distribute them. The Country claims more resources are needed in schools to improve performance even as the government tries to achieve 100% transition.
The Project Goals
– Develop a dashboard that shows the facilities to students ratio and the distribution rate of facilities in the three categories of schools in the county.
– Build a machine learning model that analyzes the performance of the school and the level of facilities in different academic years.
– Compare performance, number, and type of facilities in public and private schools.
– Outline recommendations on which facilities the county and other education stakeholders should prioritize to improve school performance and achieve 100% transition across the levels.
The Learning Outcomes
1. Recommend a check sheet that education stakeholders can use while allocating money for infrastructure development in schools.
2. Visualizations of the distribution of facilities, student population, and performance in the county.
The Tasks & Timeline
Week 1
| Week 2
| Week 3
| Week 4
|
–Weather Data Collection -Satellite image data collection. -Fishing data collection. | – Exploratory Data Analysis (EDA) -Interactive plots with Real-time data -Interactive map for the sub-counties
| -Feature Engineering – Model selection techniques Modeling | – Model Evaluation Report writing – Presentation |
Completed Project(s)
Nairobi, Kenya Chapter
Identifying Flood-Prone Regions And Assessing Impact Of Floods In Kenya Using Satellite Data
The Problem
Each year in Kenya we have the rainy seasons in March to May and also in October to December. The rains usually come with blessings for farmers but at the same time destruction when it rains too much. They lead to the destruction of infrastructure, crops, and also the displacement of people.
In this project, we aim to build a tool that identifies flood-prone regions and also how much damage the rains will cause in the selected areas. This tool will use collected satellite data to identify such regions. We shall collect data before and after the rains and then define a metric to assess the damage.
Such a tool will help the government and other agencies allocate enough resources before the rains and also allow people in the affected areas to better plan themselves before the rains start.
The Project Goals
This is more of an educative project and we hope to
1. Have a working tool that can help save lives and better prepare our community for the next rainy season
2. Create awareness on open-source data that can be used to solve problems in our community
The Learning Outcomes
1. Teach members of the community how to access open-source satellite image data, process it, and use it in other applications
2. Building of AI models that work with such data
3. Get members to feel how a project is conceptualized until its implemented
Source Code: https://github.com/OmdenaAI/omdena-kenya-floodimpact
Link to Original Project: Identifying the Impact of Desert Locust in Kenya
Kenya Chapter Leads
