A small step towards the epistemic decentralization of science: A dataset of journals and publications indexed in African Journals Online

The journal list and publications indexed in AJOL were retrieved using web scraping techniques. The metadata of the publications were retrieved using the R package ojsr (Becerra, 2024), which scrapes the metadata of the articles indexed in Open Journal Systems (OJS) pages from their head tag in html. For the journals’ information, the data were downloaded from various pages of AJOL’s website using the rvest R package (Wickham et al., 2024). Both publications’ and journals’ metadata were retrieved in February 2024.

Articles were matched against the OpenAlex (Priem et al., 2022) data dump using journal names, DOIs, and other article metadata. The process was conducted in several steps, prioritizing accuracy at each stage. Rather than maximizing the number of matches, the approach focused on ensuring accuracy. After each step, checks were performed to remove any duplicate rows from the resulting table. Matched articles were removed from the initial AJOL table so that the remaining steps were applied only to non-matched articles. In cases of duplicate matches, duplicates were removed from the matched article table and added back to the non-matched table for inclusion in the subsequent steps.

• Exact match by DOI: A total of 80,667 matches were identified.

• Exact match by journal and title: 67,191 matches were found after excluding common titles (e.g. “introduction”, “review”) to avoid false positives.

• Exact match by journal and fuzzy title match: 7,265 matches were identified. To account for inconsistencies between the databases, all titles were transformed to lowercase, and punctuation was removed. A Levenshtein distance threshold of less than 20 characters and a distance-to-title length proportion of less than 0.25 were applied to minimize errors. A sample of 100 matches was manually verified, and no errors were detected.

• Exact match by journal, year, issue, volume, and pages: 435 matches were identified in this step. Since the pages field can be inconsistent in both OpenAlex and AJOL, this step was performed last to reduce potential errors. All matches were manually checked, and 10 erroneous matches were removed.

In total, 155,558 articles were successfully matched out of 223,981 articles in AJOL. This aligns with recent studies on OpenAlex coverage. For example, Chavarro and Alperín (2024) reported that OpenAlex covers 70% of journals using the Open Journal System (OJS). Since AJOL does not record some metadata fields, matching AJOL with OpenAlex provides additional insights into the data, such as affiliation and institution analysis or collaboration patterns at different levels.

However, several important limitations related to the matching process must be acknowledged. The following three figures illustrate the percentage of articles indexed in OpenAlex based on the matching procedure described above, disaggregated by year, research field, and country. Figure 1 reveals that the coverage rates for articles published before 1995 are significantly lower than those published thereafter. It is important to contextualize this finding by noting that the volume of publications in AJOL prior to 1995 is itself quite limited, typically ranging from 20 to 40 articles per year, with a few exceptions; for instance, AJOL recorded 83, 80, and 103 publications in 1987, 1988, and 1991, respectively. The number of indexed publications began to increase in 1997 (198 publications), exceeded 1,000 in the year 2000, and consistently ranged between 10,000 and 15,000 annually from 2010 onward.

Figure 1.

Percentage of publications not matched in OpenAlex by year.

Figure 2 presents the proportion of non-matched publications by research area, bearing in mind that journals, and consequently their articles, can be assigned to up to three subject areas in the AJOL. The figure indicates that certain fields, such as History and Communication and Library Science, exhibit relatively low coverage in OpenAlex, with over 50% of their publications not indexed. However, most research areas show coverage rates between 60% and 80%. Religion is the only field with a coverage rate exceeding 80%, and only 15% of its publications are not included in OpenAlex.

Figure 2.

Percentage of publications not matched in OpenAlex by research area (AJOL journal classification).

Regarding countries, Figure 3 shows that those with the highest number of journals indexed in AJOL—namely Nigeria and South Africa, with 78,918 and 60,091 publications, respectively, compared to 21,904 from the next most productive country, Kenya—also exhibit the lowest levels of coverage in OpenAlex. In contrast, the percentage of unmatched publications for all other countries remained below 5%.

Figure 3.

Percentage of publications not matched in OpenAlex by journal country.

The final database was divided into six tables, linking publications, journals, and related metadata. Figure 4 shows the database diagram, Table pub stores publication-level information, including DOIs, titles, and bibliographic data, and connects to supporting tables for abstracts, keywords, authors, and external identifiers (e.g. OpenAlex). The source table contains journallevel metadata, such as ISSNs, country, and quality classification (e.g. JPPS status), while the source_area table associates journals with research areas.

Figure 4.

Database diagram.^②

Figure 5 shows the distribution of journals across African countries, with each country shaded based on the total number of journals hosted. The color scale on the right indicates the number of journals per country, ranging from light green (low journal count) to dark green (high journal count).

Figure 5.

Number of journals by country.

The countries with the highest numbers of journals are Nigeria (298 journals) and South Africa (105 journals). Other countries with relatively high numbers of journals include Ethiopia (53 journals), Kenya (48 journals), Ghana (38 journals), and Tanzania (36 journals). In contrast, many countries, especially in Central and Northern Africa, are home to only a few journals. For example, countries like Libya, Botswana, and Namibia host only one to three journals. Some countries show moderate levels of academic journals, with countries such as Algeria (19 journals), Zimbabwe (14 journals), and Cameroon and Uganda (13 journals) occupying the middle range. Countries such as Chad, Central African Republic, Guinea, Guinea-Bissau, Liberia, Mauritania, Mali, Namibia, Niger, Somalia, and Western Sahara do not appear on the map, as they do not have any journals indexed in AJOL.

This map highlights significant disparities in the distribution of academic journals across Africa, with a concentration of journals in only a few countries. While countries like Nigeria and South Africa are major hubs for academic publishing, many other African nations have very limited journal output, or none at all, which reflects the disparities between countries, even in a specialized source.

Figure 6 shows the distribution of journals across the JPPS quality categories and types of access. The graph on the left highlights the concentration of journals in the 1 Star, 2 Stars, and No Stars categories. This suggests that while many journals meet basic standards, fewer journals reach the highest thresholds (3 Stars), and a significant portion of journals either fail to meet star rating criteria or are inactive.

Figure 6.

Number of journals by JPPS category (left) and number of journals by access type (right).

The substantial number of Inactive Titles suggests that there are numerous journals that have ceased publishing but are still considered part of the AJOL, which highlights the role of the platform as an archive of African journals to improve its access and visibility regardless of them being active or not, differentiating it from traditional indexes, whose main objective is to index and rank active journals.

The graph on the right side of the Figure shows a higher number of diamond access journals than journals with other access types. This distribution is in line with the AJOL policy of supporting open access and free-fee publication models.

Figure 7 presents the total number of publications by year (green bars) and the relative share of publications with Digital Object Identifiers (DOI) (orange line) over time. From the 1990s onward, there is a significant increase in publications, reaching a peak between 2010 and 2020. The apparent decrease in 2024 is attributed to the data collection date, February 2024, meaning that publications released after this point were not yet included in the dataset.

Figure 7.

Publications (bar plot) and publications with DOI (orange line) by publication year.

The number of publications with DOIs follows a similar pattern to total publications, indicating the increasing adoption of DOIs as a standard identifier in scholarly publishing. This line remains aligned with the overall trend after 2000, suggesting that a significant proportion of publications have since been assigned DOIs. However, despite the growth in publications continues until 2015, there is a slight decline in the number of DOIs. To better assess the causes of this decline, the metadata in AJOL were compared with the metadata available in OpenAlex, which revealed that the drop was a cause of changes in AJOL’s reporting practices rather than a decline in the adoption of DOIs.

Figure 8 presents the distribution of publications across different research areas, illustrating the total number of publications per area. In AJOL, research areas are assigned at the journal level, meaning that the research areas of the publications correspond to the fields of the journals. As the publications’ areas depend on the journals’ classification, they are both plot together for comparison. The x-axis lists the research areas, while the y-axis represents the number of journals (green bars) of publications (orange line).

Figure 8.

Publications and journals by research area (journal area classification).

Health dominates the distribution, with over 80,000 publications and 200 journals, making it by far the largest field in terms of output. Following Health, Biology and Life Sciences, and Agriculture and Food Sciences also have significant publication numbers, though at considerably lower levels than health, each with between 30,000 and 40,000 publications. In the case of journals, Biology and Life Sciences and Humanities present the highest numbers after Health, followed by African Studies and Economics and Development. In both publications and journals, History, Philosophy, and Earth Sciences have the lowest representation, with fewer than 1,000 publications each.

The relationship between the number of journals and the number of publications is generally proportional, although fields like Agriculture & Food Science or Veterinary Science show higher publication output compared to their journal count. Other fields like Finance & Management, show a lower publication output than the number of journals. Differences in the trends of journals and publications might be due to the different publication patterns between journals. For instance, some journals in AJOL publish only one issue per year, while others publish one every few months. The number of papers per issue can also influence the relationship between the number of journals and papers.

Table 3 shows the number of publications in the AJOL according to the authors’ country. For this analysis, the matching between OpenAlex and AJOL was used and then merged with the OpenAlex data dump. Author country data were found for 94,168 publications. No normalization was applied for articles with more than one author. The two most productive countries are Nigeria (25,902) and South Africa (24,755), followed by three non-African countries: the United States (5,528), China (3,919), and India (3,615). In Africa, other countries with high productivity are Ethiopia (3,132), Ghana (2,723), Kenya (2,461), Tanzania (1,970), Morocco (1,637), Egypt (1,479), Uganda (1,410), Cameroon (1,307), Benin (1,258), Ivory Coast (1,249), and Zimbabwe (1,225). The rest of the continent, particularly the Central and Sahel regions, show lower publication outputs. Outside Africa, some countries in. Europe and the Middle East show high publication outputs, while South America and the Southeast of Asia present lower values^③.

The results highlight that African journals serve as international venues, attracting contributions from authors both within and outside the African continent. This global authorship pattern demonstrates the broad appeal and relevance of African-based journals in disseminating research across a variety of regions. Many authors from countries in the Global North, including the United States, the United Kingdom, and parts of Europe, contribute to African journals, suggesting that these journals are not only regionally focused but also engaged in the global exchange of knowledge.