How We Use Data

Last updated: 9 August 2026

Data is at the heart of Gombea.

We use it to connect people with their representatives, explain Kenya's political landscape, identify patterns in representation and make information that is spread across many different sources easier to explore.

Our approach is based on a few simple principles:

Use public information responsibly. Keep sources visible. Separate facts from analysis. Protect user privacy. Remain politically neutral.

Gombea is an independent information aggregator. We are not a government service, political party, campaign organisation or political advocacy platform.

1. Three Types of Data

It is useful to distinguish between three different types of information used by Gombea.

Public Information

Information concerning representatives, candidates, elections, political parties, electoral areas and public institutions.

This generally comes from publicly accessible sources.

Gombea-Generated Data

Information we calculate, classify or derive from public information to make it easier to understand.

Examples include political-composition statistics, age ranges and professional-background categories.

User Data

Information relating to people who use Gombea, such as account information, preferences or location when permission has been provided.

These categories are treated differently.

2. Public Political Information

Gombea brings together publicly available information about Kenya's political and public-representation landscape.

Depending on availability, this may include:

names of representatives
elected offices
counties
constituencies
wards
political-party affiliations
candidate information
election results
parliamentary activity
committee membership
public appointments
education
professional history
official biographies
public statements
relevant news coverage

This information may originate from Parliament, electoral authorities, county institutions, government publications, political parties, news organisations and other public sources.

Our Sources & Attribution page explains this in greater detail.

3. Geography Connects the Data

Political representation is geographic.

Gombea connects public information to geographic areas such as:

Kenya
↓
County
↓
Constituency
↓
Ward

This makes it possible to explore representation through a map rather than having to know the name of a particular politician.

For example, selecting a location can allow Gombea to determine the relevant:

President and Deputy President
Speakers and Deputy Speakers of Parliament
Cabinet Secretaries and other appointed national offices
Governor
Senator
Woman Representative
Member of Parliament
Member of County Assembly

depending on the geographic level being viewed.

4. Political Party Data

Political affiliation is an important part of understanding representation.

Gombea may use party information to show:

a representative's political affiliation
which party holds an elected office
party representation across an area
historical changes in affiliation
political composition across geographic areas

Party affiliation can change.

Where possible, we therefore associate political affiliation with a particular period rather than treating it as a permanent characteristic of an individual.

For example:

2022 Election — Party A
2025 — Party B

This allows historical information to remain accurate when circumstances change.

5. Political Maps

Public political information can be combined with electoral boundaries to create interactive map layers.

For example, a party-control map might show:

Kenya level

Party of each Governor

County level

Party of each MP by constituency

Constituency level

Party of each MCA by ward

Users can then explore how representation changes geographically.

Map colours indicate data categories. They do not indicate Gombea's support for a political party.

6. Representation Is Not Voter Support

This distinction is important.

If Gombea shows:

Party A — 45%

that percentage must identify what is being measured.

For example:

45% of MCA seats in the selected county

does not mean:

45% of voters support Party A.

Seat representation, election vote share, polling and public sentiment are different measurements.

Gombea aims to identify these clearly rather than combine them into misleading statistics.

7. Gender Data

Where reliable information is available, Gombea may use gender information to understand the composition of political representation.

This can help answer questions such as:

What percentage of elected representatives in an area are women?
How does representation vary between counties?
Which wards or constituencies are represented by women?
How has gender representation changed between elections?

Gender information is used to describe political representation.

It is not intended to assess the ability, suitability or performance of an individual.

Where reliable information is unavailable, it should remain Unknown rather than being inferred solely from a person's name, photograph or other unreliable indicator.

8. Age Data

Where a reliable date of birth is publicly available, Gombea may calculate a representative's age.

Age may then be grouped into ranges such as:

18–30
31–40
41–50
51–60
60+

This allows Gombea to show information such as:

average representative age
age distribution
youngest representatives
changes in age composition between elections

Age should be calculated from the underlying date rather than stored permanently as a fixed value.

Where reliable age information is unavailable, it remains Unknown.

9. Education Data

Public biographies sometimes include education histories.

Gombea may organise this information into consistent qualification categories such as:

Primary
Secondary
Certificate
Diploma
Bachelor's
Postgraduate Diploma
Master's
Doctorate
Professional Qualification
Other

We aim to retain the original education information alongside any Gombea classification.

Importantly, absence of information does not mean absence of education.

If a reliable source does not provide education information, Gombea should display Unknown rather than making an assumption.

10. Professional Background

Political representatives often have diverse careers before and during public life.

A person might have worked as a:

Teacher
↓
School Principal
↓
Trade Union Official
↓
Member of Parliament

Reducing that history to a single label such as "politician" would lose useful information.

Gombea may therefore retain individual career positions while also organising them into broader categories.

These may include:

Law & Legal Services
Education
Business & Entrepreneurship
Finance & Banking
Public Administration
Healthcare
Agriculture
Engineering & Technology
Media & Communications
NGO / Development
Trade Unions & Labour
Security & Defence
Academia & Research
Politics
Other

A person may appear in several categories during their career.

11. Primary Professional Background

For some aggregate statistics, Gombea may derive a primary pre-politics professional background.

This allows statistics to be calculated without counting the same representative several times.

For example:

Nakuru elected leadership

Business & Entrepreneurship — 31%
Education — 22%
Law — 14%
Public Administration — 12%
Other / Unknown — 21%

This is a Gombea classification derived from available career information.

It is not necessarily a professional description chosen by the representative themselves.

Where practical, the underlying employment history should remain available for users to inspect.

12. Election Data

Gombea may use historical and current election information to show:

candidates
political parties
election results
winning margins
voter turnout
historical office holders
changes in political control

During election periods, candidate information may change rapidly.

Gombea may distinguish between statuses such as:

publicly declared aspirant
party nominee
officially cleared candidate
elected
unsuccessful candidate

These statuses should be based on the best available source at the relevant time.

13. Historical Data

We do not necessarily delete political information simply because it is no longer current.

Historical information can help explain how representation changes.

Gombea may therefore retain previous:

representatives
elections
political affiliations
offices
candidate records
electoral results

This could eventually allow users to explore political maps across different election periods.

For example:

2013 → 2017 → 2022 → 2027

Historical records should be clearly distinguishable from current information.

14. News Data

Gombea may connect representatives and political areas with relevant reporting from external news organisations.

We may process information such as:

headline
publisher
publication date
article link
people mentioned
places mentioned
political organisations mentioned
topic

Gombea does not aim to reproduce complete news articles.

Users should be able to visit the original publisher for the complete story.

15. News Summaries

Where appropriate, Gombea may use automated systems to create short summaries of publicly available reporting.

A generated summary is not the original article.

It should be presented as a Gombea-generated summary and linked back to the source material.

Automated summaries can make mistakes, particularly where an article contains ambiguity or complex political context.

The original source should therefore remain available for verification.

16. Sentiment

Gombea may eventually analyse sentiment within available public information.

Possible categories might include:

media sentiment
public discussion
parliamentary discussion

Sentiment analysis is inherently approximate.

It should not be presented as an objective measure of whether a politician is "good" or "bad."

It also should not be represented as an opinion poll unless the underlying information genuinely comes from a properly conducted poll.

Where automated sentiment analysis is used, Gombea should explain what information was analysed.

17. Turning Data Into Insights

Gombea may calculate aggregate insights from underlying public data.

Examples include:

party composition
gender representation
age distribution
average age
professional-background distribution
education distribution
first-term representation
independent representation
election margins
geographic political composition

These statistics allow users to understand patterns that are difficult to see from individual records.

18. Data Tells a Story — Without Taking a Side

Gombea may translate statistics into short factual explanations.

For example:

“34% of the elected offices included in this Nairobi view are held by women.”

or:

“Five political parties hold elected offices within the selected area.”

These statements describe the underlying dataset.

They should not become political commentary.

Gombea does not use these insights to tell users whether a particular political outcome is desirable.

The data tells the story. The interpretation belongs to the user.

19. Unknown Is Valid Data

Public datasets are rarely complete.

We therefore treat Unknown as an important value.

We do not assume:

someone's age from their appearance
gender from their name
education from their job
profession from their political office
political support from browsing behaviour
current party from an old election result

It is better for Gombea to say:

Unknown

than to present an unsupported assumption as fact.

20. Data Provenance

Where practical, important information should retain a connection to its source.

Conceptually:

Source
↓
Original value
↓
Normalised value
↓
Gombea classification
↓
Gombea insight

This helps distinguish information published by an original source from information subsequently organised or calculated by Gombea.

21. Data Freshness

Political data can become stale.

A representative may:

resign
change party
lose office
be replaced
die while serving
be affected by a court decision
participate in a by-election

Gombea may therefore store information such as:

source
last checked date
verification status
effective date
historical period

Where appropriate, the interface may show this information to users.

22. Community Participation

Users may help Gombea identify potentially incorrect or outdated records.

A user report is treated as a signal for review, not as proof.

Reports may be checked against available:

official records
gazettes
electoral information
parliamentary records
county information
credible recent reporting

Only after review should the underlying record be changed.

23. Automated Processing

Because public information exists across many different formats, Gombea may use software and artificial intelligence to help:

extract information
match names
identify duplicate records
connect politicians to electoral areas
classify careers
classify education
identify news references
summarise information
calculate statistics
detect potentially stale records

Automation helps us process information at scale.

It does not eliminate the need for source verification.

24. Confidence and Verification

Some derived information may include an internal confidence level.

For example, an automated career classification might be:

Education → School Leadership

with a high confidence because the original position was:

Secondary School Principal

More ambiguous classifications may be marked for review.

Where uncertainty materially affects what users see, Gombea should prefer transparency over false precision.

25. How We Use User Data

Data about Gombea users is treated differently from public political information.

User data may be used to:

operate accounts
authenticate users
remember preferences
provide requested features
provide location-based map functionality
maintain security
understand product performance
improve Gombea
provide and measure advertising where permitted

Our Privacy Policy explains these uses in greater detail.

26. Location

With permission, Gombea may use your location to determine the political geography relevant to you.

For example:

Your location
↓
Ward
↓
Constituency
↓
County
↓
Relevant representatives

Location provided for this purpose does not mean Gombea considers that location to represent your political beliefs.

Where practical, users can search for an area manually instead.

27. Browsing Does Not Equal Political Support

This is a core Gombea data principle.

Viewing:

a politician
a political party
an election
a news story
a constituency
a political map

does not mean that you support the subject being viewed.

Gombea should not infer political beliefs merely from ordinary browsing activity.

We do not intend to create behavioural advertising profiles based on inferred political opinions from normal use of Gombea.

28. Public Profiles

Gombea may allow registered users to create public profiles.

Only information users deliberately choose to make public should appear publicly.

Private information such as:

passwords
precise device location
advertising identifiers
private account information

should not become part of a public profile.

Users should consider carefully what information they choose to publish publicly.

29. Advertising

Advertising helps fund Gombea.

Advertising systems may use information required to:

display advertisements
measure impressions
measure clicks
prevent advertising fraud
understand advertising performance

Where required, consent will be obtained before personalised advertising technologies operate.

Gombea's political-information dataset should remain separate from advertising decisions.

Payment for advertising should not alter:

political records
party maps
verification status
election results
political statistics
source information

Advertising and independent information should remain clearly distinguishable.

30. What We Do Not Want Data To Become

Gombea is not intended to become a system for deciding:

which politician is "best"
which political party people should support
how someone should vote
whether someone's age makes them suitable for office
whether someone's gender makes them suitable for office
whether someone's education makes them suitable for office
whether someone's profession makes them suitable for office

These are decisions and interpretations for individuals to make.

Our role is to organise and explain available information.

31. Our Data Principles

When making decisions about data, Gombea aims to follow these principles:

Source it

Know where important information came from.

Preserve it

Keep original information where practical before normalising it.

Structure it

Make fragmented information easier to use.

Verify it

Check important information against appropriate sources.

Date it

Political information exists within a particular point in time.

Explain it

Make statistics understandable.

Show uncertainty

Unknown is better than invented.

Protect users

Public political information and private user information are different things.

Stay neutral

Data presentation should not become political advocacy.

Let people explore

Provide the information and tools. Let users reach their own conclusions.

Our Approach

Gombea exists because useful public information is often scattered across websites, reports, databases and documents.

We bring that information together.

We connect it.

We structure it.

We make it visual.

We explain where it came from.

And then we let you explore.

We aggregate the information. You explore the picture.