LEAD. Leaders' Visits & Agreements Database

Data & Methods

What the database records, how each entry was coded, how confident we are in it, and how the releases have evolved. Variable definitions are given in English, the language of the coding sheet.

Current release

1.3Version
4,382Entries
46Country datasets
36Country profiles
5Regions
22 June 2026First published
CC BY 4.0Licence
10.5281/zenodo.22255254DOI

Release history

VersionDateCoverageWhat changed
1.32 September 202646 datasets · 4,382 entries · 5 regionsSame rows as 1.2. The machine-readable codebook is republished without coder identities; the website gained the Data & Methods page, the accessibility statement, embeddable cards and charts, and a Zenodo DOI in every citation.
1.224 August 202646 datasets · 4,382 entries · 5 regionsDataset-wide normalisation of labels, names and venues; semantic vetting passes closed; ten incidentally coded groups added as map-only datasets.
1.117 August 202636 datasets · 4,361 entries · 5 regionsSecond team-form import: 677 submitted rows, 495 accepted after de-duplication and re-orientation of inbound entries.
1.022 June 2026First public release of the database and the interactive map.

What the database records

Each row is one international engagement — an in-person visit, a phone call or a video conference — by a head of state, head of government or foreign minister of an African country, between 1990 and 2025. An engagement is described by 59 variables in eight blocks: identification of the actor, the event itself, delegation and counterparts, activities on the margins, the substance of discussions, agreements signed, the lifecycle of those agreements, and assessment, sourcing and provenance.

Public and gated fields

The interactive map, the country profiles and the downloadable data.json carry every field except four: the coders' free-text notes, the coder's identity, the evidential confidence rating and the source URLs. Those travel with the full dataset, which registered users obtain from the Archive; they are withheld from the public copy to protect coders and to keep source links from being scraped out of context.

Coding conventions

Confidence scale

A five-level ordinal scale grading the evidential basis of a record. The current release uses only the two highest levels — Explicit (3,361 records) and Implied (1,019) in the v4.2 profile; the lower three are defined for completeness and are not used.

LevelMeaning
ExplicitThe fact is directly and unambiguously stated in a source.
ImpliedStrongly supported by, but not literally stated in, the source(s).
AssumedInferred from context / standard practice without direct evidence.
GuessedLow-evidence best estimate.
DubiousRecorded but of doubtful reliability; flagged for review.

Discussion-topic vocabulary

The 23 categories observed, with the number of records in which each appears in the v4.2 profile (a record may carry several).

CategoryRecords
Diplomacy and Foreign Relations3,784
Regional Organizations and Multilateral Affairs2,537
Trade and Economic Relations1,309
Security and Defense728
International Conflicts and Crises515
Development and Aid484
Peace Negotiations and Mediation143
Symbolic and Ceremonial Issues130
Peace and Security124
Environment and Climate Change114
Financial and Monetary Issues106
Infrastructure and Connectivity91
Energy86
Health57
Terrorism and Counterterrorism54
Human Rights and Governance52
Education and Culture41
Agriculture and Food Security36
Maritime and Territorial Issues31
Migration and Human Mobility16
Technology and Innovation15
Science10
Domestic Political Matters9

Agreement families

Signed agreements are classified into four families, each with its own type field. The type vocabularies are data-derived and grow with the file.

Variable catalogue

All 59 variables in worksheet order, grouped in eight blocks. The Q-number is the questionnaire number in the column header; the short name is the machine-readable field name. Coverage and value counts are from the v4.2 profile of 4,380 records.

Block A — Identification & Actor (Q1–Q4)4 variables

Q1region

Source column: 1.) African Region? · Type: Categorical (nominal) · Coverage: complete

Macro-region of Africa under which the home actor is classified. The file spans five coded regions: the Eastern-Africa module (nineteen states), the fifteen-member ECOWAS belt in full plus Mauritania-less Western Africa (Nigeria, Ghana, Senegal, Côte d'Ivoire, Sierra Leone, Mali, Guinea, Niger, Burkina Faso, Liberia, Togo, Benin, The Gambia, Guinea-Bissau and Cabo Verde) and Sudan (Northern Africa), and South Africa (Southern Africa — the fourth region, opened in v4.0); v4.2 adds ten stub groups created by re-orienting team-coded inbound visits into outbound records of the visiting African state (Egypt, Libya, Tunisia, Morocco, Algeria, Angola, Democratic Republic of the Congo, Equatorial Guinea, Botswana, Namibia — opening Central Africa as the fifth region and extending Northern and Southern Africa); these stub groups are not systematically built.

Eastern Africa2,739Western Africa1,204Northern Africa300Southern Africa132Central Africa5

Q2home_country

Source column: 2.) Home Country? · Type: Categorical (nominal) · Coverage: complete

Sovereign state (or de facto state) the travelling / engaging leader represents (the sending actor). Forty-six home actors across five regions, thirty-six of them systematically coded; the rest are incidental groups created by re-orienting inbound visits.

Ethiopia480Kenya409Sudan290Nigeria242Zambia176Somalia175Tanzania169Uganda167

Q3actor_title

Source column: 3.) Title? (King, President, Prime Minister, Foreign Minister, Defense Minister, Finance Minister, etc.) · Type: Categorical (nominal) · Coverage: complete

Official office held by the home actor at the time of the engagement. Free office label drawn from an open list; 27 distinct values clustering on heads of state/government and senior ministers.

President2,826Foreign Minister843Prime Minister440Deputy Prime Minister & Foreign Minister45Defence Minister40Deputy President40Finance Minister35Vice-President25

Note: Top 14 of 27 distinct office labels shown; the tail comprises single- to triple-count ministerial and military-chief titles (State Minister of Finance, Chief of Defence Staff, Minister of Petroleum/Trade/Mining/Investment, Army/Coast-Guard Commander, Special Envoy, etc.).

Q4leader_name

Source column: 4.) Leader's Name? · Type: Open text · Coverage: complete

Personal name of the home actor. High-cardinality identifier (579 distinct individuals across the period).

Values: 579 distinct. Most frequent: Uhuru Kenyatta (170), Yoweri Kaguta Museveni (122), William Ruto (111), Paul Kagame (99), Omar al-Bashir (98), Muhammadu Buhari (95), Meles Zenawi (91), Abiy Ahmed (87), Samia Suluhu Hassan (70), Ismail Omar Guelleh (69), Robert Mugabe (69), Hakainde Hichilema (66).

Block B — Engagement Event (Q5–Q9)5 variables

Q5meeting_type

Source column: 5.) Meeting Type? · Type: Categorical (nominal) · Coverage: complete

Modality of the engagement: an in-person trip, a telephone call, or a video conference.

Visit4,229Video Conference76Phone Call75

Q6country_visited

Source column: 6.) Country Visited/Called/Conferenced? · Type: Categorical (nominal) · Coverage: complete

Counterpart / host country of the engagement (destination of a visit, or the country of the interlocutor for a call or video conference).

Values: 130 distinct. Most frequent: United States of America (1,416), China (214), Ethiopia (210), France (127), Kenya (121), Tanzania (113), Egypt (101), South Africa (98), United Kingdom (98), Russian Federation (96), Uganda (89), United Arab Emirates (75).

Note: Country labels use long-form controlled values (e.g. “Russian Federation”, “Türkiye”, “United Kingdom of Great Britain and Northern Ireland”).

Q7city_visited

Source column: 7.) City Visited/Called/Conferenced? · Type: Categorical (nominal) · Coverage: 4,357 valid / 23 missing

City in which the engagement took place (for calls, typically the interlocutor’s seat).

Values: 300 distinct. Most frequent: New York (1,214), Addis Ababa (204), Beijing (189), Washington (179), Nairobi (115), Paris (102), London (74), Cairo (73), Kampala (72), Kigali (66), Djibouti City (63), Khartoum (59).

Q8start_date

Source column: 8.) Trip/Call/Conference Start Date? · Type: Date (MM/DD/YYYY) · Coverage: complete

First day of the engagement. Stored as text in US MM/DD/YYYY order.

Values: Range 01/01/1990 → 12/30/2025. All 4,380 values parse as valid dates.

Q9end_date

Source column: 9.) Trip/Call/Conference End Date? · Type: Date (MM/DD/YYYY) · Coverage: complete

Last day of the engagement (equal to the start date for single-day events).

Values: Range 01/01/1990 → 12/30/2025. All 4,380 values parse as valid dates.

Block C — Delegation & Counterparts (Q10–Q20)11 variables

Q10hogs_cotravel

Source column: 10.) HoG/S Co-Travel? · Type: Binary (Yes / No) · Coverage: complete

Whether another head of government/state from the home country travelled together with the actor (joint travel).

No4,360Yes20

Q11meet_host_hogs

Source column: 11.) Meet with Host Country HoG/S? · Type: Binary (Yes / No) · Coverage: complete

Whether the actor met the host country’s head of government/state.

No3,097Yes1,283

Q12host_hogs_name

Source column: 12.) If Yes, Name? (Title First Name Last Name) · Type: Open text · Coverage: 1,283 valid / 3,097 missing

Name (with title) of the host head of government/state met. Conditional on Q11 = Yes (skip otherwise).

Values: 314 distinct. Frequent counterparts: President Xi Jinping (71), President Yoweri Kaguta Museveni (46), President Vladimir Putin (38), President Ismail Omar Guelleh (38), President Emmanuel Macron (37), President Abdel Fattah el-Sisi (34), President Isaias Afwerki (34), President Omar al-Bashir (29), President Hu Jintao (29), President Uhuru Kenyatta (28), Prime Minister Abiy Ahmed (28).

Q13meet_host_subhogs

Source column: 13.) Met With Host-country Sub-HoG/S Equivalent? · Type: Binary (Yes / No) · Coverage: complete

Whether the actor met a sub-head-of-government counterpart of the host (e.g. a foreign or other senior minister) rather than the HoG/S.

No4,038Yes342

Q14host_subhogs_name

Source column: 14.) If Yes, Name? (Title First Name Last Name) · Type: Open text · Coverage: 342 valid / 4,038 missing

Name (with title) of the host-country sub-HoG/S counterpart met. Conditional on Q13 = Yes.

Values: 232 distinct, e.g. Foreign Minister Sergey Lavrov (12), Foreign Minister Wang Yi (10), Prime Minister Narendra Modi (8), Secretary of State Antony Blinken (7), Secretary of State Marco Rubio (5).

Q15summit_attend

Source column: 15.) Attend a Multilateral Summit or Event? · Type: Binary (Yes / No) · Coverage: complete

Whether the engagement included attendance at a multilateral (heads-level) summit or event.

Yes2,648No1,732

Q16summit_name

Source column: 16.) If Yes, Name of Summit or Event? · Type: Open text · Coverage: 2,648 valid / 1,732 missing

Name of the multilateral summit/event attended. Conditional on Q15 = Yes.

Values: 1,068 distinct, e.g. United Nations Millennium Summit (16), the UN General Assembly General-Debate sessions (up to 13 each), Presidential Inauguration (12), AU Summit (10), FOCAC Beijing Summit 2006 (9).

Q17min_summit_attend

Source column: 17.) Attend a Multilateral Ministerial Event or Summit? · Type: Binary (Yes / No) · Coverage: complete

Whether the engagement included a multilateral event at ministerial (sub-heads) level.

No4,284Yes96

Q18min_summit_name

Source column: 18.) If Yes, Name of Summit or Event? · Type: Open text · Coverage: 96 valid / 4,284 missing

Name of the ministerial multilateral event. Conditional on Q17 = Yes.

Values: 81 distinct, e.g. the trilateral foreign-ministers’ talks on the Grand Ethiopian Renaissance Dam (Ethiopia–Sudan–Egypt), BRICS Foreign Ministers’ Meeting, IGAD/AU ministerial meetings.

Q19nonhost_hogs_meet

Source column: 19.) Meet HOG/S from Non-host Country? · Type: Binary (Yes / No) · Coverage: complete

Whether the actor met a head of government/state of a third country (not the host), typically on the sidelines of a multilateral event.

No4,175Yes205

Q20nonhost_hogs_names

Source column: 20.) If Yes, Name(s)? (Title First Name Last Name; List all) · Type: Open text · Coverage: 205 valid / 4,175 missing

Name(s) of third-country HoG/S met. Multiple names may be listed in one cell. Conditional on Q19 = Yes.

Values: 152 distinct strings; several cells list more than one leader (separated by “;”). Frequent: US President Bill Clinton (6), President Xi Jinping (5), Prime Minister Abiy Ahmed (4).

Block D — Activities on the Margins (Q21–Q25)5 variables

Q21public_address

Source column: 21.) Public Address? · Type: Binary (Yes / No) · Coverage: complete

Whether the actor delivered a public address / speech during the engagement.

No3,877Yes503

Q22cultural_site_visit

Source column: 22.) Visit/Address Cultural/Historical/Religious Site or Event? · Type: Binary (Yes / No) · Coverage: complete

Whether the engagement included a visit to, or address at, a cultural, historical or religious site or event.

No4,325Yes55

Q23business_leaders_meet

Source column: 23.) Meet with Business Leaders? · Type: Binary (Yes / No) · Coverage: complete

Whether the actor met business / private-sector leaders.

No4,350Yes30

Q24io_leader_meet

Source column: 24.) Meet with Leader of Int'l Organization? · Type: Binary (Yes / No) · Coverage: complete

Whether the actor met the head of an international organisation.

No4,244Yes136

Q25io_leader_name

Source column: 25.) If Yes, Name? (Title First Name Last Name) · Type: Open text · Coverage: 136 valid / 4,244 missing

Name (with title/organisation) of the international-organisation leader met. Conditional on Q24 = Yes.

Values: 91 distinct, e.g. UN Secretaries-General Kofi Annan (11) / Ban Ki-moon / António Guterres, IMF & World Bank officials, IGAD Executive Secretary.

Block E — Substance of Discussions (Q26)1 variables

Q26discussion_topics

Source column: 26.) Category of Discussion Topic(s)? · Type: Multi-select (comma-separated) · Coverage: complete

One or more thematic categories describing what was discussed. Stored as a comma-separated list drawn from a controlled vocabulary of 23 categories (full list and frequencies in §6.1).

Values: 525 distinct combinations of the 23 base categories. Most frequent single category: Diplomacy and Foreign Relations (present in 3,784 records); most frequent combination: “Regional Organizations and Multilateral Affairs, Diplomacy and Foreign Relations” (880).

Block F — Agreements Signed (Q27–Q43)17 variables

Q27agreement_signed

Source column: 27.) Signed an Agreement? · Type: Binary (Yes / No) · Coverage: complete

Master gate: whether any agreement was signed during the engagement. Opens the four agreement-family blocks (Q28–Q42).

No4,158Yes222

Q28agr_hlp_flag

Source column: 28.) High-level Politics (treaty-level) Agreement? · Type: Binary (Yes / No) · Coverage: complete

Whether a high-level political (treaty-level) instrument was signed — peace, alliance and boundary treaties, community accession/establishment treaties and founding conventions, and formal diplomatic-relations instruments.

No4,304Yes76

Q29agr_hlp_type

Source column: 29.) If Yes, What Type of High-level Politics (treaty-level) Agreement? · Type: Categorical (nominal) · Coverage: 76 valid / 4,304 missing

Sub-type of the high-level political instrument. Conditional on Q28 = Yes. The type vocabulary is data-derived and richer than the original three treaty types.

Peace Treaty33Establishment Treaty12Alliance Treaty12Joint Communiqué10Accession Treaty4Diplomatic Relations Protocol3Boundary Treaty2

Q30agr_hlp_title

Source column: 30.) Title of Agreement? · Type: Open text · Coverage: 40 valid / 4,340 missing

Official / descriptive title of the high-level political instrument (e.g. “Algiers Agreement (Eritrea–Ethiopia)”, “Pretoria Agreement”, “R-ARCSS”). Conditional on Q28 = Yes.

Q31agr_hlp_notes

Source column: 31.) Specific Notes? · Type: Open text · Coverage: 47 valid / 4,333 missing

Free-text notes on the high-level political instrument (scope, signing context, dates).

Q32agr_econ_flag

Source column: 32.) Economics-related Agreement? · Type: Binary (Yes / No) · Coverage: complete

Whether an economics-related agreement was signed.

No4,305Yes75

Q33agr_econ_type

Source column: 33.) If Yes, What Type of Economics-related Agreement? · Type: Categorical (nominal) · Coverage: 75 valid / 4,305 missing

Sub-type of the economics agreement. Conditional on Q32 = Yes.

Economic Partnership Agreement26Investment Treaty15Trade Agreement13Development Finance / Loan Agreement8Energy / Economic Cooperation Agreement5Investment Agreement4Trade and Investment Agreement2Tax Treaty2

Q34agr_econ_title

Source column: 34.) Title of Agreement? · Type: Open text · Coverage: 57 valid / 4,323 missing

Title of the economics agreement (e.g. BITs, lines of credit, LAPSSET Corridor agreements, TFTA). Conditional on Q32 = Yes.

Q35agr_econ_notes

Source column: 35.) Specific Notes? · Type: Open text · Coverage: 58 valid / 4,322 missing

Free-text notes on the economics agreement (value, parties, verification source).

Q36agr_func_flag

Source column: 36.) Functional Cooperation-related Agreement? · Type: Binary (Yes / No) · Coverage: complete

Whether a functional-cooperation agreement was signed.

No4,323Yes57

Q37agr_func_type

Source column: 37.) If Yes, What Type of Functional-cooperation Agreement? · Type: Categorical (nominal) · Coverage: 57 valid / 4,323 missing

Sub-type of the functional-cooperation agreement. Conditional on Q36 = Yes.

Environmental Agreement12Framework/Partnership Agreement11Memorandum of Understanding7Migration Agreement6Consular Agreement5Cooperation Agreement5Defence Cooperation Agreement4Labour Mobility MoU3

Q38agr_func_title

Source column: 38.) Title of Agreement? · Type: Open text · Coverage: 45 valid / 4,335 missing

Title of the functional-cooperation agreement. Conditional on Q36 = Yes.

Q39agr_func_notes

Source column: 39.) Specific Notes? · Type: Open text · Coverage: 43 valid / 4,337 missing

Free-text notes on the functional-cooperation agreement.

Q40agr_tech_flag

Source column: 40.) Technical or Administrative-related Agreement? · Type: Binary (Yes / No) · Coverage: complete

Whether a technical/administrative agreement was signed.

No4,332Yes48

Q41agr_tech_type

Source column: 41.) If Yes, What Type of Technical or Administrative-related Agreement? · Type: Categorical (nominal) · Coverage: 48 valid / 4,332 missing

Sub-type of the technical/administrative agreement. Conditional on Q40 = Yes.

Memoranda of Understanding30Administrative Agreement12Technical Cooperation Agreement5Memorandum of Understanding1

Note: “Memoranda of Understanding” (30) and “Memorandum of Understanding” (1) are number variants of the same label.

Q42agr_tech_title

Source column: 42.) Title of Agreement? · Type: Open text · Coverage: 37 valid / 4,343 missing

Title of the technical/administrative agreement. Conditional on Q40 = Yes.

Q43agr_tech_notes

Source column: 43.) Specific Notes? · Type: Open text · Coverage: 35 valid / 4,345 missing

Free-text notes on the technical/administrative agreement.

Block G — Agreement Lifecycle / Status (Q44–Q49)6 variables

Q44status_provisional

Source column: 44.) Still in the Provisional Status? · Type: Categorical (nominal) · Coverage: complete

Whether a signed agreement remains provisional (not yet entered into force). “N/A” marks records with no agreement.

N/A4,270No107Yes3

Q45provisional_sig_date

Source column: 45.) If Yes, Signature Day? · Type: Date (MM/DD/YYYY) · Coverage: 10 valid / 4,370 missing

Signature date of a provisional agreement (MM/DD/YYYY). Sparsely populated.

Q46status_ratified

Source column: 46.) Ratified? · Type: Categorical (nominal) · Coverage: complete

Whether the agreement has been ratified. “N/A” marks records with no agreement.

N/A4,273No102Yes5

Q47ratification_date

Source column: 47.) If Yes, Ratification Date? · Type: Date (MM/DD/YYYY) · Coverage: 5 valid / 4,375 missing

Ratification date (MM/DD/YYYY). Conditional on Q46 = Yes; sparsely populated.

Q48status_terminated

Source column: 48.) Terminated? · Type: Categorical (nominal) · Coverage: complete

Whether the agreement has been terminated. “N/A” marks records with no agreement.

N/A4,271No108Yes1

Q49termination_date

Source column: 49.) If Yes, Termination Date? · Type: Date (MM/DD/YYYY) · Coverage: 1 valid / 4,379 missing

Termination date (MM/DD/YYYY). One value present.

Block H — Assessment, Sourcing & Provenance (Q50–Q59)10 variables

Q50strategic_importance

Source column: 50.) Strategic Importance? (High, Medium, Low) · Type: Categorical (ordinal) · Coverage: complete

Coder’s ordinal assessment of the strategic significance of the engagement.

Medium3,129High939Low312

Q51source_link_1

Source column: 51.) Source Link 1? · Type: Open text · Coverage: complete

Primary source URL documenting the engagement. Always populated.

Values: 2,790 distinct URLs. Most-cited domains: en.wikipedia.org (760), digitallibrary.un.org (521), u.afp.com (226), gadebate.un.org (225), allafrica.com (193), ask.un.org (183), advance.lexis.com (158), history.state.gov (73), mfaethiopia.blog (73).

Q52source_link_2

Source column: 52.) Source Link 2? · Type: Open text · Coverage: 1,554 valid / 2,826 missing

Second corroborating source URL (optional).

Q53source_link_3

Source column: 53.) Source Link 3? · Type: Open text · Coverage: 310 valid / 4,070 missing

Third corroborating source URL (optional).

Q54source_link_4

Source column: 54.) Source Link 4? · Type: Open text · Coverage: 96 valid / 4,284 missing

Fourth corroborating source URL (optional).

Q55source_link_5

Source column: 55.) Source Link 5? · Type: Open text · Coverage: 34 valid / 4,346 missing

Fifth corroborating source URL (optional).

Q56source_document

Source column: 56.) Source Document? (pdf, doc, docx, mp3, mp4, jpeg, png, etc.) · Type: Open text · Coverage: 0 valid / 4,380 missing

Field reserved for an attached/archived source document (file type).

Note: Entirely empty in this release (0 / 4,380 populated).

Q57confidence_level

Source column: 57.) Confidence Level? (Explicit, Implied, Assumed, Guessed, Dubious) · Type: Categorical (ordinal) · Coverage: complete

Coder’s confidence in the record, on a five-level evidential scale (see §6.3). In this release only the two highest levels are used.

Explicit3,361Implied1,019

Note: The scale defines five levels (Explicit → Dubious); the current release contains only Explicit and Implied — the lower three are defined but unused.

Q58coder_name

Source column: 58.) Name of Coder? · Type: Categorical (nominal) · Coverage: complete

Name of the coder who entered the record (provenance / inter-coder tracking).

Note: Coder identities are withheld from the public codebook; the field travels only with the full (gated) dataset.

Q59additional_info

Source column: 59.) Additional Info? · Type: Open text · Coverage: complete

Free-text narrative summary of the engagement (one paragraph per record). Always populated and near-unique to each record; functions as the human-readable description of the row.

Values: 4,357 distinct narratives across 4,380 records.

Known limitations

Download the data

The public dataset is archived on Zenodo: download release 1.3 (data, machine-readable codebook and licence in one bundle), or use the all-versions DOI, which always resolves to the newest release. The full row-level dataset with sources remains available to registered users in the Archive.

Document control

This page is generated from codebook v4.2, itself produced by automated column-by-column profiling of the source workbook (SHA-256 9ec7b9c1…). The codebook is re-issued whenever the source file changes; the live database currently holds 4,382 entries.

Cite and reuse

The database is licensed CC BY 4.0. Cite it with the reference on the About page; work that uses it is listed under Outputs.