SayPro datasets

SayPro Dataset Review (20 Sets)

#	Dataset Title	Type	Segment Dimensions	Coverage	Usefulness	Notes
1	National Skills Training Cohort 2023	Demographic, Behavioral	Age, Gender, Completion, Dropouts	All 9 provinces	High	Strong course-level granularity
2	Youth Unemployment Program Outcomes	Demographic, Behavioral	Age 18–35, Education, Job placement	Gauteng, KZN, EC	High	Limited rural comparison
3	SayPro Online Platform Analytics (2024–25)	Behavioral	Device type, Session frequency, Drop-off points	National	High	Good for digital access patterns
4	Cultural Programme Attendance Logs	Demographic, Geographic	Age group, Province, Event type	6 provinces	Medium	Demographics incomplete for 2022
5	Legislative Impact Feedback – SCRR	Demographic, Geographic	Community type, Age, Political district	40 districts	Medium	Qualitative-heavy, requires NLP
6	Women in Informal Economy Report	Demographic, Behavioral	Gender, Employment status, Training uptake	EC, WC, GP	High	Focused but very relevant
7	SayPro Survey – Learning Preferences	Behavioral	Format preferences, Barriers to access	4,000 respondents	High	Self-reported, useful for design
8	Radio + USSD Awareness Pilot (2023)	Geographic, Behavioral	Region, Access medium, Conversion	Limpopo, NW, FS	High	Useful for low-bandwidth outreach
9	Neftaly Cultural Tracker	Demographic, Geographic	Ethnicity, Language, Attendance	5 provinces	Medium	Still being standardized
10	Technical & Vocational Training Impact	Demographic	Age, Prior education, Employment outcomes	National	High	Post-course follow-up included
11	Refugee & Migrant Engagement Baseline	Demographic, Geographic	Country of origin, Age, Language	GP, WC	Medium	Needs behavioral overlays
12	Literacy & Numeracy Programme Data	Demographic, Behavioral	Age group, Assessment scores, Device type	7 provinces	High	Good for tracking baseline shifts
13	Trainer Observation Reports	Behavioral	Participation, Motivation, Group dynamics	150+ sessions	Medium	Qualitative, needs structured coding
14	Digital Content Engagement (Video, SCORM)	Behavioral	Click-through, Completion, Feedback	All LMS users	High	Automated data source
15	Community Partner Feedback Reports	Geographic	Area-specific needs, Cultural barriers	Rural sites	Medium	Unstructured, some segment data missing
16	Mobile App Usage Logs (SayPro App)	Behavioral	Feature use, Time of day, Retention	12,000 users	High	AI-ready data
17	Gender-Based Violence Response Pilot	Demographic, Behavioral	Gender, Experience type, Support uptake	EC, WC	Medium	Sensitive data; ethics protocols required
18	Early Childhood Development (ECD) Trainers	Demographic	Age, Qualification, Regional presence	FS, NW, MP	Medium	Not yet digitized
19	Employment Readiness Bootcamp Surveys	Demographic, Behavioral	Age, Job readiness, Pre/post comparison	600 participants	High	Includes impact scores
20	WhatsApp Chatbot Log Analysis	Behavioral	Query type, Sentiment, Re-engagement rate	All segments	High	Real-time feedback insights

---

📌 Key Observations

• High behavioral segmentation potential exists in digital platform and chatbot datasets.
• Geographic coverage is uneven, with underrepresentation in the Northern Cape and rural KwaZulu-Natal.
• Demographic data is strongest in formal training programs but weakest in community outreach and cultural work.
• Many qualitative datasets need standardization for integration with dashboards or AI tools.

---

✅ Next Steps

• Prioritize integration of top 10 high-value datasets into the GPT-based segmentation system.
• Develop data cleaning scripts and metadata standards for under-structured sources.
• Launch a 2025 universal data collection framework requiring gender, age, region, and program type.
• Automate behavioral tagging across online and in-person activities.