scpmicro/adnan-project-fieldwork-dashboard
Adnan Project Fieldwork Dashboard v2.4 — Root-Only Deployment
Conducted by Sheekar Research.
A bilingual English/Arabic Flask dashboard for Excel and manual fieldwork updates, workbook-derived targets, completion analysis, remaining workload, QC monitoring, management insights and full-screen presentation.
Core operating rule
No quantitative target is hard-coded in the application. The active target and all supporting quota structures are extracted from the uploaded master workbook, including:
- region and city targets;
- age-group and SEC quotas;
- product-mention targets;
- city × age and city × SEC matrices;
- region × product targets;
- Focus Group and IDI targets;
- QC targets.
Uploading a different valid master workbook replaces the active target structure. Operational progress workbooks use the current active target plan.
User workflow
- Choose English or العربية.
- Sign in for data-management access.
- Upload Excel/CSV or use Manual Entry.
- Review the live dashboard.
- Open Presentation Mode for executive delivery.
- Use Reset actual data before beginning a new operational dataset, or upload a new master workbook with Replace selected.
Default administrative access
- Username:
admin - Password:
123
For a public deployment, replace these through Hugging Face Secrets:
FIELDWORK_ADMIN_USERNAMEFIELDWORK_ADMIN_PASSWORDFIELDWORK_SECRET_KEY
Hugging Face configuration
Use Docker, CPU Basic (Free) and port 7860.
Recommended variables:
FIELDWORK_DATA_DIR=/data/fieldwork_quota_intelligenceFIELDWORK_DEMO_MODE=0FIELDWORK_COOKIE_SECURE=1
Attach a storage bucket to /data if uploaded data and approvals must survive rebuilds and restarts.
Main capabilities
- dynamic target and quota extraction from the active workbook;
- native Adnan workbook priority plus adaptive English/Arabic Excel mapping;
- administrator Mapping Review for ambiguous sheets and headers;
- Excel/CSV upload and manual entry;
- independent Interview and Product Mention analytical bases;
- Product 1/Product 2, multi-response and long-format product support;
- conservative product ontology with unmapped-value reporting;
- critical/warning/information validation severity;
- Interview-ID QC deduplication and explicit non-verifiable QC states;
- row-level and aggregate Focus Group / IDI actuals;
- recruitment, seating, QC and transcript gap tracking;
- dynamic completed, remaining and completion-rate KPIs;
- city, age, SEC, product, QC and qualitative analysis;
- automatic management interpretations;
- bilingual English/Arabic LTR/RTL interface;
- executive light theme with optional dark mode;
- Reset Actual Data and Start New Workbook controls;
- full-screen Presentation Mode with Completed and Remaining Work at the end;
- printable bilingual final report;
- import diagnostics and auditable human approval.
Import decision model
- Imported: high-confidence schema and valid records.
- Imported with warnings: analysis continues; warnings remain visible.
- Mapping required: administrator confirms uncertain sheet/column mappings.
- Critical hold: impossible arithmetic or critical structural defects remain excluded until approval.
See EXCEL_SCHEMA_COMPATIBILITY.md for supported formats and analytical-base rules.
Support
Dashboard support contact: Scpmicro@gmail.com. The email is intentionally omitted from Presentation Mode.
Root-only Hugging Face deployment
This release is intentionally folder-free at repository level. Upload all files in this package directly to the root of the Space.
app_assets.zip contains the application templates, CSS, JavaScript, Plotly runtime, workbook templates and reference files. At startup, app.py verifies and securely extracts the archive to a temporary runtime directory before loading app_core.py.
Required root files:
README.mdDockerfilerequirements.txtapp.pyapp_core.pyadaptive_excel.pyapp_assets.zip.gitattributes.gitignore.dockerignore
No templates/, static/, data/ or runtime_data/ folders need to be uploaded manually.
