A
Background
Latar Belakang
A distribution company managing multiple product lines, transporters, and regional sales territories. The business generated substantial operational data daily through their ERP system and field operations, but that data existed in silos — reports were produced manually, consolidated by hand, and delivered too late to influence decisions.
Sebuah perusahaan distribusi yang mengelola berbagai lini produk, transporter, dan wilayah penjualan regional. Bisnis ini menghasilkan data operasional yang besar setiap hari melalui sistem ERP dan operasi lapangan, namun data tersebut tersimpan di silo — laporan diproduksi secara manual, dikonsolidasi dengan tangan, dan disampaikan terlambat untuk mempengaruhi keputusan.
B
Operational Problem
Masalah Operasional
Supply chain performance was invisible. Management relied on weekly manual consolidation of spreadsheets, WhatsApp summaries, and email reports — with no unified view of delivery costs, sales trends, or transporter performance.
Performa rantai pasokan tidak terlihat. Manajemen bergantung pada konsolidasi manual spreadsheet mingguan, ringkasan WhatsApp, dan laporan email — tanpa tampilan terpadu untuk biaya pengiriman, tren penjualan, atau kinerja transporter.
C
Existing Workflow
Alur Kerja Saat Ini
Data came from multiple Excel files updated at different frequencies. By the time trends were consolidated and visible, problems had already escalated into financial losses or operational disruptions.
Data berasal dari berbagai file Excel yang diperbarui pada frekuensi berbeda. Saat tren selesai dikonsolidasi dan terlihat, masalah sudah berkembang menjadi kerugian finansial atau gangguan operasional.
1
Operations team pulls daily export from ERP into Excel files
2
Each department maintains its own consolidated sheet independently
3
Weekly WhatsApp reports sent by coordinators to management
4
Management requests specific analysis ad hoc — hours or days later
5
Decisions made based on data that is 3–7 days old
1
Tim operasional mengambil ekspor harian dari ERP ke file Excel
2
Setiap departemen mengelola sheet konsolidasi masing-masing secara mandiri
3
Laporan WhatsApp mingguan dikirim oleh koordinator ke manajemen
4
Manajemen meminta analisis spesifik secara ad hoc — beberapa jam atau hari kemudian
5
Keputusan dibuat berdasarkan data yang sudah 3–7 hari yang lalu
D
Bottleneck & Risk
Bottleneck & Risiko
By the time trends became visible, the window for corrective action had already closed. A transporter with consistently increasing delivery costs would only be flagged during monthly review. A region with declining sales performance would surface only when the quarterly target was missed.
Saat tren mulai terlihat, jendela waktu untuk tindakan korektif sudah tertutup. Transporter dengan biaya pengiriman yang terus meningkat baru ditandai saat review bulanan. Wilayah dengan performa penjualan yang menurun baru terlihat ketika target kuartal sudah terlewatkan.
Poor-performing transporters continued without detection. High-cost routes were unnoticed until period-end. Sales anomalies were discovered in retrospect, too late for corrective action.
Transporter berkinerja buruk terus beroperasi tanpa terdeteksi. Rute berbiaya tinggi tidak terlihat hingga akhir periode. Anomali penjualan baru ditemukan secara retrospektif, terlambat untuk tindakan korektif.
E
Why the Existing System Failed
Mengapa Sistem Lama Gagal
The existing system was designed for data recording, not operational visibility. ERP exports were structured for accounting, not for operational decisions. Nobody owned the responsibility of making data visible at the speed operations required. The tools existed — the workflow to surface insights in real time did not.
Sistem yang ada dirancang untuk pencatatan data, bukan visibilitas operasional. Ekspor ERP terstruktur untuk keperluan akuntansi, bukan untuk keputusan operasional. Tidak ada yang memiliki tanggung jawab untuk membuat data terlihat secepat yang dibutuhkan operasional. Alatnya ada — alur kerjanya untuk menampilkan insight secara real-time tidak.
F
Solution Approach
Pendekatan Solusi
Rather than modifying the existing ERP or building a heavy new system, exported operational data was processed through Python and presented in a decision-focused analytical dashboard built around control points — not generic charts.
Daripada memodifikasi ERP yang ada atau membangun sistem baru yang berat, data operasional yang diekspor diproses melalui Python dan disajikan dalam dashboard analitik yang berfokus pada keputusan — dibangun di sekitar titik kontrol, bukan grafik generik.
G
System Architecture
Arsitektur Sistem
Python-based analytics dashboard with sales trend analysis, delivery cost monitoring per route, transporter performance comparison, and operational KPI tracking refreshed from automated data exports.
Dashboard analitik berbasis Python dengan analisis tren penjualan, pemantauan biaya pengiriman per rute, perbandingan kinerja transporter, dan pelacakan KPI operasional yang diperbarui dari ekspor data otomatis.
A Python-based data pipeline reads from automated ERP exports dropped to a shared folder. Pandas handles cleaning, transformation, and aggregation. Streamlit renders an interactive dashboard with Plotly charts. No database required — the pipeline runs on schedule and rebuilds views from the latest exports.
Pipeline data berbasis Python membaca dari ekspor ERP otomatis yang disimpan ke folder bersama. Pandas menangani pembersihan, transformasi, dan agregasi data. Streamlit merender dashboard interaktif dengan grafik Plotly. Tidak diperlukan database — pipeline berjalan terjadwal dan membangun ulang tampilan dari ekspor terbaru.
H
Technologies Used
Teknologi yang Digunakan
Python
Streamlit
Pandas
Plotly
I
Workflow Visualization
Visualisasi Alur Kerja
01
Data Export
Ekspor Data
ERP exports pushed automatically to shared folder on schedule
Ekspor ERP otomatis dikirim ke folder bersama sesuai jadwal
02
Processing
Pemrosesan
Python pipeline reads, cleans, and aggregates operational metrics
Pipeline Python membaca, membersihkan, dan mengagregasi metrik operasional
03
Dashboard
Dashboard
Streamlit renders interactive views — cost analysis, KPIs, trends
Streamlit merender tampilan interaktif — analisis biaya, KPI, tren
04
Decision
Keputusan
Management accesses current data without waiting for manual reports
Manajemen mengakses data terkini tanpa menunggu laporan manual
J
Operational Impact
Dampak Operasional
Metric
Metrik
Before
Sebelum
After
Sesudah
Detection Speed
Kecepatan Deteksi
5–7 days
Same day
Report Preparation
Persiapan Laporan
4–6 hours manual
Automated refresh
Transporter Visibility
Visibilitas Transporter
Monthly review
Real-time comparison
Decision Lag
Keterlambatan Keputusan
1 week avg
Next morning
Delivery cost anomalies detectable within hours instead of weeks
Transporter performance comparable across time periods and routes
Management reporting reduced from manual weekly consolidation to automated daily refresh
Decision-making grounded in current operational data
Anomali biaya pengiriman dapat terdeteksi dalam hitungan jam, bukan minggu
Kinerja transporter dapat dibandingkan lintas periode waktu dan rute
Pelaporan manajemen berubah dari konsolidasi manual mingguan ke pembaruan harian otomatis
Pengambilan keputusan berbasis data operasional terkini
K
Future Development
Pengembangan ke Depan
Direct ERP API integration to eliminate the export-drop step. Predictive cost anomaly alerts via threshold rules. Automated email digest for management. Mobile-optimized view for field operations access.
Integrasi langsung dengan API ERP untuk menghilangkan langkah ekspor-drop. Peringatan anomali biaya prediktif melalui aturan threshold. Ringkasan email otomatis untuk manajemen. Tampilan yang dioptimalkan untuk mobile bagi akses operasi lapangan.