Understanding internet penetration within communities is crucial for equitable development, but countries like Togo lack comprehensive data on connectivity demand, especially in households, hospitals and schools. Accurate insights are essential for effective policymaking and resource allocation, and machine learning offers useful tools in this realm, to predict internet usage patterns and bridge data gaps.
The objective of this competition is to develop a machine learning solution to predict fiber to the home (FTTH) internet uptake across various segments in Togo using socio-demographic and geographic datasets. The goal is to create accurate and generalisable models to help the government improve connectivity, allocate resources efficiently, and foster digital inclusivity.
Your solutions will address socio-economic challenges and drive sustainable development in Togo by enabling better policies and infrastructure investments, and narrowing the digital divide. Identifying institutions that need internet access will improve healthcare and education, fostering a more connected, empowered, and inclusive Togolese society.
The top 20 participants will have the possibility to be interviewed by Togo's Ministry of the Digital Economy and Transformation. And the 10 performers participants will work on a two-year contract for the Datalab, which will be a physical space dedicated to modelling data in Togo with a view to transforming information into concrete actions as solutions to the various challenges in all areas of Togo.
A project of the Ministry of the Digital Economy and Transformation through the operational arm of Agence Togo Digital.
The Togolese Republic is firmly committed to modernising its economy, with the aim of making Togo a hub for business and innovation by 2025. This vision has materialised in the adoption of a government roadmap of forty-two (42) priority projects and reforms (three quarters of which have a digital component) and a strategy for the digital transformation of Togo by 2025, ‘Stratégie Togo Digital 2025’, which integrate all the country's digitalisation projects into a coherent whole.
The mission of the Ministry of the Digital Economy and Transformation in Togo is to promote the country's digital development, with projects in the fields of e-government, digital identity and digital regional planning.
The ministry's remit and objectives are to develop the digital economy and digital transformation in Togo. Its activities focus on areas such as entrepreneurship, infrastructure, innovation, broadband internet and telecommunications.
The main projects include e-ID Togo, e-government and digital regional planning, which have enabled Togo to post encouraging key figures in the digital sector, including a high penetration rate of fixed and mobile internet.
To support the implementation of the Togo digital 2025 strategy, the Agence Togo Digital (ATD) is the entity responsible for supporting the Togolese administration in its digitisation process, while promoting innovation. It is expected to digitise 75% of administrative procedures, in line with the objectives of the government's 2025 roadmap. It is working actively with the private sector (start-ups, independent developers, digital service providers, etc.) and government departments to achieve this objective.
The error measure for this competition is the error Area Under the Curve.
For each row in the dataset, the submission files must contain 2 columns: ID and Target.
Where 1 indicates that there is Internet access and 0 indicates that there is no Internet access.
Your submission file should look like this:
ID Target
ID_002V80RTY6 0.78
ID_006GHNSAZX 0.34
1st place: FCFA 1,500,000
2nd place: FCFA 1,000,000
3rd place: FCFA 500,000
Possibility of a 2-year contract for the 10 candidates to work in the Datalab after an interview phase for 20 best candidates with the Ministry of the Digital Economy and Transformation.
This challenge is open only to Togolese citizens. Winners will need to provide prove of residence.
There are 3 000 Zindi points available. You can have more information here: Zindi points
The competition starts on 10 May 2024.
The competition ends on 23 June 2024 at 23:59 PM.
We reserve the right to update the contest timeline if necessary.
This challenge is open only to Togolese citizens.
Teams and collaboration
You may enter this competition as an individual.
Multiple accounts per user are not permitted, nor is collaboration or membership between users. Individuals and their contributions from multiple accounts will be immediately disqualified from the platform.
Code must not be shared privately. Any code shared must be made available to all participants in the competition via the platform (i.e. on the discussion forums).
Datasets and packages
The solution must use only publicly available open-source software.
You may only use the datasets provided as part of this competition. Machine learning tools such as AutoML are not allowed.
You may use pre-trained models, provided they are accessible to all.
The data used in this contest is the exclusive property of Zindi and the contest organiser. You may not transmit, duplicate, publish, redistribute or otherwise provide or make available the contest data to any party not participating in the contest (this includes uploading the data to a public site such as Kaggle or GitHub). You may upload, store and work with the data on any cloud platform such as Google Colab, AWS or similar, as long as 1) the data remain private and 2) this does not contravene Zindi's usage rules.
You must notify Zindi immediately upon becoming aware of any unauthorised transmission or access to Contest Data, and work with Zindi to rectify any unauthorised transmission or access.
Your solution must not infringe the rights of any third party and you must be legally entitled to assign to Zindi ownership of all copyrights in the code of the winning solution.
Submissions and winning
You can make a maximum of 10 submissions per day.
You can make a maximum of 200 submissions in total in this competition.
Before the end of the competition, you must choose 2 submissions to be judged for the private ranking. If you don't make a choice, your two best entries in the public ranking will be used for the private ranking.
During the competition, your best public score will be displayed, regardless of the entries you have selected. At the end of the competition, your best private score on the two selected submissions will be displayed.
Zindi maintains a public and a private ranking for each competition. The public ranking comprises around 40% of all test data. During the competition, the public leaderboard ranks the submitted solutions according to the accuracy score they have achieved. At the close of the competition, the private ranking, which covers the remaining 60% of the test data set, will be made public and constitutes the final ranking of the competition.
Please note that to be taken into account, your submission must first be processed. If your submission fails during the processing stage, it will not be counted or scored, nor will it count towards your daily submission limit. If you encounter any problems with your submission file, the best thing to do is to ask for advice on the contest discussion forum.
If you're in the top 5 when the ranking closes, we'll send you an e-mail asking for your code. On receipt of the e-mail, you will have 2 hours to reply and submit your code, following the submission guidelines detailed below.If you do not respond, you will be disqualified.
If your solution is one of the winning solutions, you will be required to submit the code of your winning solution to us for verification, and you thereby agree to assign to Zindi all worldwide copyrights relating to that winning solution.
If two solutions obtain identical scores on the leaderboard, the date and time of submission will be decisive (the older solution will win).
If the error metric requires probabilities to be submitted, don't set thresholds (or round off your probabilities) to improve your ranking. To ensure that the customer receives the best solution, Zindi will need the raw probabilities.This will allow customers to set thresholds according to their own needs.
You acknowledge and agree that Zindi may, without obligation to do so, remove or disqualify any person, team or account if Zindi believes that such person, team or account is in violation of these rules.Participation in this competition implies acceptance of the official competition rules.
Zindi is committed to providing valuable solutions to its customers and partners. To this end, we reserve the right to disqualify your submission for reasons of usability or value. This includes, but is not limited to, the use of data leakage or any other practice that we consider to compromise the inherent value of your solution.
Please consult the FAQ and terms of use for additional rules that may apply to this competition. We reserve the right to update these rules at any time.
Reproducibility of submitted code
Data standards:
Consequences of breaking any rules of the competition or submission guidelines:
Monitoring of submissions
Join the largest network for
data scientists and AI builders