Ellie Boyce, Amber Dennison, Orlagh McSorley, Nancy Pilkington, and Rowanne Smith with Fozia Bora and Yahya Birt

Introduction

Working over several months, a team of five students analysed the photographic archive held by the Somali Village charity (est. 2025), as part of its Somali Village in Colonial Bradford (SVCB) project, alongside community historian Yahya Birt and Professor Fozia Bora. The archive we examined consisted of 31 relevant photos from Marseilles (2) and Bradford (29) from the 1904 tour. The wider archive also includes other photos from later tours that this particular Somali troupe undertook in Germany in 1905‒6 and 1907, with some overlapping personnel. Through this examination, we aimed to get an accurate number of individuals, or a more precise estimate. Each week, we laid the photos out, studying them carefully to match people across different photos. Initially, progress was quite slow, but, as we got used to the photos, it became easier to identify individuals, and we began to recognise them.

We only had ethical approval to research the archive, but not to publish any images as part of our research, because the sensitive nature of these images of ethnographic performers (Collomb, 2011), especially of the children, makes them inappropriate for publication. The goal of the Somali Village charity is to make the archive available to future researchers upon request if deemed necessary for their projects, but the images will not be in the public domain without careful consideration and supervision from the Somali community.

This report summarises our aims, research process, findings and conclusions.

Aims

Our main aim was to establish how many individuals performed: partly to counter misconceptions from contemporary advertising, which claimed there were 100 Somalis ‒ a popular way of marketing shows (Birt, 2026), and partly to give a more accurate depiction of performers’ lives during their stay in Bradford.

We also aimed to identify individuals to determine if any performers also travelled to later shows in Europe by comparing photographs from Bradford with those from a particular stop in Mannheim, where the troupe performed on one of its later tours (Birt, 2025). Tracing individuals’ trajectories would demonstrate how ethnographic performance became a way of life and was understood by performers as a career, rather than forced exploitation (Bora, 2025), and to highlight the variety of experiences of the practice, as some performers remained in Somaliland upon returning from Bradford, while others returned to Europe to tour again (Birt, 2026; Warsame, 2021).

Finally, we aimed to establish any family connections or groups to explore how the Somali Village functioned and was staged.

Process

Drawing on Professor Hassan Ugail’s pioneering facial recognition technology, which we used for a preliminary identification of matches between individuals from the Bradford photographs, we familiarised ourselves with the photographs and began to identify recurring individuals. This technology, developed by Professor Ugail and his team at the University of Bradford’s Centre for Visual Computing and Intelligent Systems, which operates within academic ethical standards, allowed us to match the same or possibly related individuals across multiple photographs in this archive, even those of relatively poor and grainy quality (Ugail, 2025). The technology works by scanning each image for faces, then converting each face into a unique numerical “barcode” based on set features such as the distances between the eyes, nose, mouth, cheekbones, jawline and similar features. To check whether two photographs show the same person, it compares those barcodes and produces a match score out of 100 ‒ the closer the score to 100, the more likely the faces belong to the same individual, with 70 being the cut-off point. However, while this technology helped us initially, we found that there was no substitute for working together as a group to recognise individuals through features unavailable to Prof. Ugail’s facial recognition technology, such as hairstyles, dress and ornaments, which proved important in identifying individuals.

Initially, photographs from Bradford and Marseilles, France, were used to identify individuals. These individuals were then given codes, separating them into three groups: children, women and men, and numbered according to the order they were identified in. The codes were written on sticky labels and placed on the A3 prints of the photographs, blown up from postcard size (approx. 13cm by 9cm) to make identification of details easier. Some photographs were of poor quality and were replaced with higher-quality copies when they became available to facilitate identification.

We found that, between sessions, labels could be lost or displaced, which necessitated that we keep a database to track which photographs individuals appeared in. We logged the codes given to individuals on a spreadsheet, listing the photographs they appeared in, and gave descriptions of their appearance to facilitate future identification and reduce the chance of double-counting.

We first worked on identifying the children, as they were the most traceable group, because they were well-photographed and often shot together in proximity, which made identification easier. However, some postcards of the children’s facial features had been altered by being touched up by hand to give them an exotic or even threatening appearance, which made identification more challenging. We categorised the children into girls, boys and infants, using non-facial features like height, clothing, hair and staging, alongside facial ones to guide our selections.

Adult women were the second most traceable group as they were significantly fewer in number, and were often posed with children, or for group portraits, making their facial features the centre of the composition and thus easier to identify. The one major group shot of women with children shows them washing clothes.

It was tempting to use the family portraits as evidence of family connections, especially as the set-ups sometimes differed from stereotypical expectations, for example, one family group of adults was captioned as a “native family” on the postcard, which was interesting, as no children were grouped with them. Ultimately, however, the photographs are poor guides as to the genuine relationships between individuals. One exception was that Professor Ugail’s technology did identify one individual as highly likely to be Sultan Ali’s brother, standing to his left in a group shot entitled “Sultan Ali and his bodyguard”. Despite this, considering the significance of staged family groups was important for our research because it gave us insight into how the Somali Village was created and marketed.

The three staged-families were as follows:

The first group is labelled as “Somali Village. The family of the Native Doctor.” It features two women, both holding children, one a girl and the other an infant. Next to them stands a girl who was often photographed in Bradford wearing a silver matrimonial and dowry necklace known as the Laasin. Sat on the ground in front of them are two girls and a boy. The group is placed in front of the Chakim or Doctor’s display hut.
The second group is labelled as “Somali Native Family, Bradford Exhibition.” It features five adults, three men and two women stood around a brazier with a painted backdrop featuring palm trees.
The third group is not labelled as a family, although it is posed as one that stands between the Chakim’s hut and the saddler’s station. Although it was originally taken in Bradford, this is one of the postcards that was later issued in Germany in 1906 and labelled as “Abyssinisches Dorf” (Abyssinian Village). A man stands with two women. A third, older woman sits holding a small girl. Three girls are placed in front, two standing and one sitting. Behind them is the painted backdrop with palm trees. Notably, two girls featured in Family 1 reappear here, staged as part of Family 3.

Finally, we identified men, who were much harder to identify, as there were significantly more of them. They were often photographed as groups in landscape, action shots, involved in performances like battle reenactments or playing cards. These scenes were often edited as well, making the depictions less reliable. Our identifications were thus less certain than those of women and children. Two men, however, could be identified by name, Sultan Ali al-Urfa and Daniel Mesango, who feature in multiple photographs. As the leader of the troupe, Sultan Ali al-Urfa was photographed in close proximity, often accompanied by his ‘guard’, one of whom, as suggested above, is likely to have been his brother. The roles assigned to some men, like the Imam, the Saddler and the Teacher, meant they were photographed performing their associated roles, for example, respectively leading prayer, selling items at the Saddler’s station, or teaching children in the Village.

During the write-up stage, we conducted an analysis of the data we had collected in order to draw some broad conclusions, which we set out below.

Outcomes

We were able to identify 58 individuals: 27 children, 11 women and 20 men. The children were further identified as 15 boys, 11 girls and 1 infant.

The total nearly matches the figure given in French reporting on the troupe’s arrival in Marseille from Somaliland – 57 Somalis (Birt, 2026). The fifty-eighth person has been identified by the SVCB research team as a non-Somali who was also part of the Village – the Sierra Leonean Daniel Mesango, who was most likely hired by the impresario Victor Bamberger to train the Somalis in ethnographic performance, as this was their first tour. He had previously worked with Bamberger as part of his Ashanti Village in the 1890s. The assumption is that Mesango joined the troupe in France or later in England. The fact that all members of the group have been positively identified is a remarkable result. It cements the conclusion that the Somali troupe in Bradford was among the best visually documented among ethnographic troupes as a whole.

When considering family connections, besides the three staged families discussed above, we identified four potential relationships in addition to Prof. Ugail’s identification of Sultan Ali’s brother: two different sets of potential siblings photographed together two and three times, and two mother-child pairings. These potential relations may be particularly useful when considering later photos in Mannheim or elsewhere in Germany to see if family units returned together to Europe for later tours. We ran out of time on the project to complete a comparison of the Mannheim and Bradford photographs to identify individuals who travelled again to Europe, but we understand this work will be continued as part of the SVCB project. Our initial findings were that there is a strong match between two women who were in Bradford in 1904, who appear again in Mannheim in 1907.

Our analysis of the archive also shows that girls were more photographed than boys, that children were more photographed than adults, and that men were more photographed than women. Of the ten most photographed individuals, nine were children – with Sultan Ali al-Urfa, the troupe leader, being the only adult in the list with four photographs. This focus on children may have been an outcome of commercial imperatives in the print industry in generating postcard sales.

Group	n	Avg photos per person
Girls	11	2.82
Boys	15	2.47
Men	20	1.75
Women	11	1.36
Infants	1	1.00

The adult gender imbalance in the troupe at roughly 2:1 male to female is striking for what was presented as a community or village. This reflects other imperatives at play other than family life in staging the ethnographic village. Most important were the mock fights and dances, which were performed by the men, and the world of work as represented in the market stalls, as well as functionaries like the Imam and the Teacher.

We hypothesise that, given the number of Somali men (19), women (11) and children (27), there were several smaller family groups, and several unmarried men or married men who travelled to Europe without their spouses or children at home. According to Warsame (2021: 91), it was common for Somali ethnographic troupes to be composed of the troupe leader’s family members, close relations and allies in his clan, in this case, Sultan Ali’s.

Ultimately, our research on the photographic archive has been particularly fruitful in further exploring the experiences of individuals in the Somali Village. Working on this part of the project has been particularly rewarding for us as students, as we were able to help and contribute to original research as part of this project.

References

Birt, Y. 2025. A Timeline of Bradford’s Somali Village. Somali Village. [Online]. [Accessed 20^th May 2026]. Available from: A Timeline of Bradford’s Somali Village – Somali Village

Birt, Y. 2026. A Somali Village in Colonial Bradford- An Introduction. Somali Village. [Online]. [Accessed 20^th May 2026]. Available from: A Somali Village in Colonial Bradford – An Introduction – Somali Village

Bora, F. 2025. A Somali Village in Colonial Bradford – A Project. Somali Village. [Online]. [Accessed 24^th May 2026]. Available from: https://somalivillage.org.uk/a-somali-village-in-colonial-bradford/

Collomb, G. 2011. “Galibis” in Paris!. In: Blanchard, P., Boëtsch, G. and Jacomijn Snoep, N (eds.) Human Zoos: The Invention of the Savage. Paris: Musée du Quai Branly. p.287.

Ugail, H. 2025. Unpublished research note on the use of face‑recognition technology in archival picture research. Bradford: Centre for Visual Computing and Intelligent Systems, University of Bradford.

Warsame, B. 2021. 3. A Brief History of Staging Somali Ethnographic Performing Troupes in Europe, 1885–1930. In: Demski, D. and Czarnecka, D. eds. Staged Otherness: Ethnic Shows in Central and Eastern Europe, 1850–1939. [Online]. Budapest, Hungary: Central European University Press, pp. 77-100. [Accessed 9^th October 2025]. Available from: https://doi.org/10.1515/9789633864401-004

Ellie Boyce, Amber Dennison, Orlagh McSorley, Nancy Pilkington, and Rowanne Smith produced this work as an output of their module FOAH2002 Student Research Partnerships, part of the BA in Liberal Arts at the University of Leeds, during a placement with the Somali Village in Colonial Bradford project in the academic year 2025–26.