Skip to main content
Ramin Khavarzadeh DATA · ANALYSIS · PERSPECTIVE
فا
Data LAB

Data lab

Indices we build from public data, what looked unusual versus its own baseline this week, and relationships we tested among series — including rejects, not only hits.

INDICES

This week’s proprietary indices

Week starting ۱۲ مهر · computed from 43 Weekly series · Methodology

Consumer Pressure Consumer Pressure
Experimental
28 -16 vs last week
Week starting 27 Sep 2 of 5 components worsened

How hard are households looking for liquidity and credit?
Demand Pulse Demand Pulse
Experimental
19 -30 vs last week
Week starting 27 Sep 1 of 4 Component rose

How strong is the appetite for large durable goods?

Each index is 0–100: this week’s percentile rank among 52 past weeks. 50 means a typical week; 100 means the strongest week in this window. All are descriptive experiments — not forecasts. Consumer Pressure · Demand Pulse

WHAT CHANGED

What changed this week?

Series that, in the latest complete week, moved at least 1.5 standard deviations from their prior 12-week mean — 8 of 35

  1. Search Search for “rent house” −20% z -4.8
  2. Search Search for “visa” −30% z -3.1
  3. Search Search for “laptop” +13% z 2.6
  4. Search Search for “fridge” −15% z -2.2
  5. derived Durable goods basket −7% z -2.1
  6. Search Search for “TV” −9% z -2.0
  7. Search Search for “migration” −24% z -1.8
  8. Search Search for “buy gold” +23% z 1.5
SIGNALS WE FOUND

Signals we found Experimental

Lagged cross-correlation, corrected for every pair and every lag tested (217 Test)، With out-of-sample validation. Correlation, not causation.

Weak signals and relationships not found
BRAND SEARCH

Iranian brand search interest

Each brand’s figure is relative to that brand’s own highest week in the past 12 months (100 = its own peak search week) — not a sentiment score, only relative search volume. Brands are compared only to their own baseline, not to each other, because each group is normalised in a separate Google request.

#BrandSegment12-week trendThis weekWeekly change
1 Choopan 12-month low Food 0 -100.0٪
2 Delpazir Food 5 +66.7٪
3 Mahan Air Airlines 2 -33.3٪
4 Hyperstar Chain retail 2 -33.3٪
5 Saipa Auto 6 -25.0٪
6 Iran Air Airlines 6 -25.0٪
7 Shahrvand Chain retail 12 +20.0٪
8 Namava Media 4 -20.0٪
9 Alibaba App 4 -20.0٪
10 Bank Mellat Bank 10 -16.7٪
11 Filimo Media 34 -12.8٪
12 Golrang Food 9 -10.0٪
13 Snapp App 30 -9.1٪
14 Bank Melli Bank 13 +8.3٪
15 Irancell Telecom 12 -7.7٪
16 Ofogh Kourosh Chain retail 25 -7.4٪
17 Hamrah-e Aval Telecom 13 -7.1٪
18 Kaleh Food 43 -6.5٪
19 Torob App 86 -4.4٪
20 Mihan Food 44 -4.3٪
21 Bank Saman Bank 57 -3.4٪
22 Divar App 82 -2.4٪
23 Bank Pasargad Bank 80 +1.3٪
24 Tapsi 12-month high App 3 0.0٪
25 Iran Khodro Auto 7 0.0٪
26 Bank Saderat Bank 4 0.0٪
27 Rightel Telecom 50 0.0٪
28 Refah 12-month low Chain retail 3 0.0٪
29 Basalam App 28 0.0٪
30 Tak Macaron 12-month low Food 1 0.0٪
OVERLAY

Two series overlaid, with lag

Both series are converted to z-scores; the first is shifted by the optimal lag so co-movement is visible.

METHOD Each index is built from public series with thresholds set in advance. Relationships are reported with correction for all pairs and lags tested, and correlation never means causation. Full methodology

TOOL همین آزمون رابطه را روی دو سری زمانی خودتان اجرا کنید؛ داده فقط در مرورگر شما پردازش می‌شود. آزمایشگاه رابطه با دادهٔ شما