AVAILABLE FOR WORK
Meida Istiqomah
AI & Automation Solution Architect, GIS & Data Specialist
5+ years of experience as a GIS, Remote Sensing & Data Specialist. Now focused on building AI-powered automation and AI orchestrator agents that reduce manual work to make it faster and more reliable.
Turning Manual Work Into Automated Systems
Started as a GIS Analyst. Now I research, build, and automate — hunting down bottlenecks in manual workflows and turning them into systems with minimal human touch: faster, more reliable, more efficient.
Career Journey
Nov 2025 · Present
AI & Automation Solution Architect @ Chronicle Cemetery Software
- Build end-to-end automation systems integrating AI/ML models across geospatial data, headstone imagery, scanned documents, and records from multiple databases
- Cut manual processing time by 30–35% across records cleaning, headstone transcription, bookscan, and QC workflows
- Identify bottlenecks in the Project department's workflows and drive R&D to prototype high-impact automation tools
- Train the Project department on AI tools and workflows to boost adoption
- Translate complex technical requirements into actionable solution designs aligned with business goals
Sep 2024 · Oct 2025
GIS Project & Support Team Lead @ Chronicle Cemetery Software
- Oversaw end-to-end GIS delivery for 3 new clients/month, 99% QC accuracy
- Redesigned GIS workflows, improving team efficiency by 25% without compromising quality
- 2 Python automation scripts/month replacing manual, repeatable tasks
- Managed post-delivery client support, diagnosing and resolving 10–15 GIS issues per month
- Trained and onboarded new GIS team members on standardized workflows, QC procedures, and tooling
Oct 2022 · Aug 2024
GIS Analyst → GIS Support Specialist @ Chronicle Cemetery Software
- Converted multi-format legacy data (CAD, Excel, Access, drone imagery) into structured GIS databases
- Conducted rigorous QC on all project datasets to verify spatial accuracy and completeness before delivery
- Processed drone imagery (Agisoft Photoscan, Pix4D) into high-resolution orthomosaics
- Managed a high-volume support queue, diagnosing GIS issues and delivering customized spatial solutions
- Documented troubleshooting processes into a centralized knowledge base to speed up team issue resolution
Earlier Roles
- Urban Planner Expert Assistant, DPU PR Kab. Natuna (1:5,000 RDTR mapping, ArcGIS Pro + SPOT 6 imagery)
- Land Surveyor & Data Entry, Bantul Land & Spatial Planning Office (field survey & land registry)
- Staff Intern & Practicum Assistant, Bappeda Sleman & UGM Remote Sensing & GIS Diploma
2021
Diploma III, Remote Sensing & GIS @ Universitas Gadjah Mada
- Remote sensing image processing, spatial analysis, and GIS database design
- Practicum assistant & internship at Bappeda Sleman during studies
2022
Data Science @ Digital Talent Scholarship, Kementerian Komunikasi dan Informatika RI
- Government-sponsored data science training program
Selected Work
Case studies of ongoing & completed projects. This list keeps growing.
Kalimantan Wildfire Mapping & Analysis
End-to-end remote sensing pipeline: from raw satellite data to publication-ready visualization & quantitative validation based on ASEAN/CIFOR standards.
(August 2026)
(≥2 satellite sensors)
compared (2015/2019/2023/2026)
(dNBR, 3 sample sites)
Hotspot Timelapse: August 2026
NASA FIRMS data (VIIRS + MODIS), filtered by confidence level, cross-validated between sensors via spatial buffer analysis (not just coordinate rounding). 7-day rolling-window visualization follows Copernicus EMS/EFFIS (Europe) & Sentinel Asia/ASMC (Asia) cartographic standards.
- Confidence filtering + cross-sensor dedupe
- Scale bar, north arrow, province boundaries, legend
- ColorBrewer YlOrRd color scheme (research-validated, colorblind-safe)
Historical Comparison: 2015 vs 2019 vs 2023 vs 2026
Is this wildfire crisis a recurring pattern? Compared using the same sensor combination (MODIS + VIIRS_SNPP) across all four periods for a fair comparison. NOAA-20 has only been available since 2017/2018, so it was excluded from the historical analysis.
- August 2026: 108,429 hotspots, the highest of the 4 periods
- August 2023 was actually the quietest of the four (27,611) — lower than 2019, even though a moderate El Niño was already underway that August (ONI ~1.0–1.3, per NOAA)
- ENSO state alone doesn't explain the pattern: 2015 and 2026 had similar El Niño strength in August (ONI ~1.4) yet very different hotspot counts, while 2019 — ENSO-neutral, no El Niño at all — still out-hotspotted 2023
- Geographic cluster pattern consistently recurs every year
- FIRMS Standard Processing archive data for 2015/2019/2023
| Province | 2015 | 2019 | 2023 | 2026 | ||||
|---|---|---|---|---|---|---|---|---|
| total | high-conf. | total | high-conf. | total | high-conf. | total | high-conf. | |
| Kalimantan Barat | 25,924 | 4,453 | 16,363 | 2,860 | 14,236 | 2,771 | 46,823 | 9,130 |
| Kalimantan Tengah | 33,828 | 5,332 | 12,347 | 1,606 | 4,525 | 793 | 38,013 | 6,562 |
| Kalimantan Timur | 9,805 | 1,618 | 2,613 | 323 | 2,624 | 358 | 10,787 | 2,095 |
| Kalimantan Selatan | 5,820 | 752 | 2,360 | 286 | 4,675 | 612 | 8,780 | 1,429 |
| Kalimantan Utara | 3,596 | 593 | 796 | 123 | 485 | 43 | 1,716 | 209 |
Hotspot counts, August only, matched sensors (VIIRS-SNPP + MODIS) across all four years for a fair comparison. "High-confidence" is FIRMS's own per-detection quality flag — a stricter subset of "total" that's less likely to include noise (e.g. sun glint, sensor artifacts), so it's a good sanity check on the raw count.
Recurring Hotspot Zones, 2015–2026
Small-multiples to see whether the worst fire clusters appear in the same geographic location every year, indicating a recurring risk area (likely linked to peatland).
- The southwestern coastal belt of West Kalimantan shows up as a dense cluster in all four years, 2015, 2019, 2023, and 2026 alike
- The Central/South Kalimantan border area is also a recurring hotspot in three of the four years (2015, 2019, 2026), but was comparatively quiet in 2023
- Both zones sit on Kalimantan's major peatland — consistent with peat fires being harder to fully put out and more likely to reignite on the same land
- The rest of Kalimantan stays comparatively quiet in every period — activity concentrates in the same places rather than shifting around randomly
Visual Validation: Sentinel-2 & dNBR
Hotspots alone aren't enough. Some turned out to be false positives from mining sites, not vegetation fires. Validated with 10m-resolution Sentinel-2 optical imagery and calculated dNBR (delta Normalized Burn Ratio) — an index that measures how much vegetation condition changed before vs. after a fire, from satellite images — following the official Guideline on Burned Area Mapping and Estimation in Southeast Asia (ASEAN Secretariat & CIFOR, 2025).
- NBR = (NIR − SWIR) / (NIR + SWIR), Sentinel-2 bands B08/B12
- Severity classification + burned area estimate (hectares)
- Provincial hotspot ranking (own FIRMS data) matches independent news/BPBD reports for Aug 2026 — absolute counts differ by methodology, but the relative order lines up
⚠ Data is indicative, for learning/portfolio purposes, not an official government report. Sources: NASA FIRMS, Copernicus Sentinel-2 (via Copernicus Data Space Ecosystem), geoBoundaries, NOAA CPC Oceanic Niño Index (ENSO state). dNBR thresholds follow Key & Benson (2006), not yet fully ground-truth validated per Chapter 5 of the Guideline on Burned Area Mapping and Estimation in Southeast Asia (ASEAN Secretariat & CIFOR, 2025).
← Back to project listTools & Methodology
- Geospatial Intelligence
- Spatial Data Analysis
- Spatial Data Creation
- Programming for Spatial
- Cartography
- Database Management
- QGIS
- ArcGIS Pro
- Remote Sensing
- Claude Code / Gemini CLI / OpenCode / Codex
- MCP Server Development
- Multi-Agent Orchestration
- Computer Vision (YOLO)
- Vision-Language Models
- AI API Integration
- Prompt Engineering
- LLM Integration
Let's Connect
Open to discussing GIS, remote sensing, AI automation, or project collaboration.