01 / ACTIONS & ACTIVITIES
Scenario-based movement
Scope human activities, gestures, routines, workflows, interactions, or other observable sequences relevant to the agreed model task.
Commissioned video data for AI training
NDPC REGISTERED · NDPC/DCP/13596
Registered with the Nigeria Data Protection Commission. Certificate available on request. See our Compliance page and Privacy Policy.
African video data for models that need action in context.
BSG DataWorks scopes scenario-based video-data projects around the actions, settings, participants, camera conditions, labels, and quality criteria your model must handle. We prepare commissioned video data to the project brief, making consent, provenance, temporal annotation, and delivery questions visible before production begins.
Project scope
Commission video data around the temporal task—not a generic footage catalogue.
We begin with the model's action or sequence problem and confirm the practical scope before capture or preparation starts. The project can define people or objects in frame, activities, setting, devices, viewpoints, timing, labels, quality checks, and delivery requirements.
02 / SETTINGS & SPEAKERS
Daily-life and presentation contexts
Define household, commerce, workplace, presentation, or other project settings, including any on-screen speakers and the permitted visual context.
03 / CAPTURE VARIATION
Multi-angle and device-aware capture
Set the camera positions, devices, framing, lighting, duration, movement, viewpoints, and edge cases that matter for the intended deployment context.
Computer-vision use cases
Video data designed around what the system must interpret over time.
Tell us the activity, sequence, event, scene transition, or on-screen interaction your model needs to handle. We use those facts to shape a collection or preparation brief instead of assuming one footage set fits every video task.
Action and activity recognition
Define the actions, boundaries, variations, occlusions, camera viewpoints, and temporal labels needed for the target recognition task.
Event and workflow understanding
Scope the sequence of events, roles, objects, setting, timing, and annotation rules needed to interpret an agreed real-world process.
Speaker and presentation video
Prepare scoped video examples around on-screen speakers, presentations, gestures, framing, background, audio needs, and permitted usage conditions.
Temporal QA and evaluation
Prepare held-out clips, agreed temporal guidance, review criteria, and documentation to help the buyer evaluate the scoped video-data route against its own requirements.
Technical and quality requirements
Specify the video conditions that make a collection usable.
A strong video-data brief explains what happens on screen, when it happens, how the footage is captured, which variations matter, how labels are defined, and what evidence accompanies delivery.
Capture plan
- Actions, events, routines, settings, or interaction categories
- Participants, objects, locations, and inclusion or exclusion rules
- Camera devices, framing, viewpoints, lighting, resolution, and duration
- Target variation, edge cases, and sampling considerations
Temporal annotation and metadata
- Action, event, tracking, transcription, or other label schema
- Clip, frame, timestamp, or segment-boundary requirements
- Written annotation guidance and issue-escalation process
- Metadata fields relevant to the buyer's model and evaluation needs
Quality assurance
- Video and audio usability checks agreed for the project
- Temporal-label review and acceptance-criteria process
- Sampling, feedback, correction, and version-control approach
- Clear treatment of out-of-scope or unusable material
Consent and provenance
- Source and collection-route documentation appropriate to the scope
- Written consent workflow when identifiable contributors are involved
- Early flagging of biometric, child-participant, sensitive-setting, or re-identification questions
- Buyer-facing provenance and data-handling notes agreed for delivery
From scenario brief to delivery
A practical route for commissioned video-data work.
We use the temporal task and deployment context to make capture, preparation, annotation, quality, and governance decisions visible before production work begins.
Define
Clarify the model task, activities or events, participants, visual setting, audio needs, exclusions, and the temporal variation the project requires.
Design
Agree the capture rules, devices and framing, label schema, timestamp or segment treatment, metadata, quality thresholds, and evidence requirements.
Prepare
Collect or organise the scoped footage, then annotate and review it using the agreed guidance and quality process.
Deliver
Provide the video files, labels, metadata, and agreed documentation in the required structure and delivery route.
Frequently asked questions
Questions about African video data.
Do you offer a ready-made video-data catalogue?
No. Most video-data work is commissioned around the buyer's actions, scenarios, capture conditions, temporal annotation requirements, and delivery needs. We confirm the practical scope before committing to a project.
Can a project include identifiable people or children?
Potentially, but this is a special project question. The brief should explain why identifiable participants are needed, their role, how the footage will be used, and what safeguards are required. We use written consent workflows where applicable and flag biometric, child-participant, sensitive-setting, and re-identification questions early; buyers should obtain independent professional advice for their own legal and governance obligations.
Can you work with footage we already have?
Yes. If you already hold source footage, we can discuss a preparation, annotation, quality-assurance, and delivery scope around an agreed schema and review process. See the Annotation and Labelling Services page.
Can the collection cover a particular African market or setting?
Tell us the market, setting, participants, activities, capture conditions, and delivery requirement. We assess the brief and confirm what is practical before making a commitment; this page does not claim universal market coverage.
Can the project include speech or transcripts?
Yes, where this is within the agreed scope and the participation, data-use, and quality requirements are clear. If the project needs dedicated voice data, see the African Speech Data page for the related route.
Start with the temporal task
Start an African video-data brief.
Send the outline you have. We will help turn the action, scenario, capture context, and quality requirements into the right next conversation.
A useful first email includes
- Intended model or product use and action, event, or sequence task
- Activities, settings, participants, objects, speakers, or other categories to include
- Markets, locations, camera conditions, duration, audio, and capture requirements
- Target volume, video format, metadata, labels, and quality requirements
- Rules for devices, angles, lighting, framing, temporal boundaries, variation, or exclusions
- Rights, consent, governance, delivery timing, and procurement context, if known
If you do not have all of this yet, send what you have. We will help clarify the remaining questions on a call.
Prefer to write directly? dataservices@bsgdataworks.com
WhatsApp only: +234 806 790 4903