A Bloomberg-style climate intelligence terminal that turns the El Niño / La Niña (ENSO) cycle into actionable commodity and sector positioning, with the causal rigor to tell which links are real and which are spurious.
Live Demo · The Moat · Screenshots · 14 Pages · Methodology · API · Resources · Run It · Deploy · FAQ
The ENSO Macro Risk Desk is a production-grade, open source Python dashboard that answers one question for a commodity or macro analyst and for a climate aware portfolio manager: when the El Niño–Southern Oscillation (ENSO) cycle shifts, what commodity and sector exposure should you reposition, and which of those links survive causal testing?
It ingests canonical ENSO data directly from NOAA CPC (the Oceanic Niño Index), ERSSTv5 sea surface temperature grids, the World Bank Pink Sheet commodity database, and IMD 0.25° gridded rainfall area-weighted across 1901–2024. It then runs a dual model forecasting engine (SARIMA plus a PyTorch LSTM) and a causal inference engine (Granger causality, Convergent Cross Mapping, and phase-randomized surrogate significance) across fourteen interactive pages, in a dark, data-dense terminal UI. No API keys are required to start.
Every weight, threshold and known limit is written down in
docs/METHODOLOGY.md, rendered in-app as page 09 and
enforced against the code by a test, so the published document cannot drift from
what actually runs. The same caches are served as JSON at /api, so a claim
on a page can be checked rather than screenshotted.
The product philosophy is describe → prescribe: every region and commodity ends in a positioning view (constructive, cautious or watch, plus a swing catalyst and a risk), not just a chart. Those stances are computed, not typed, and gated by the causal verdict, which is why the desk will say no trade when that is what the evidence supports.
Who it's for: commodity and macro research analysts, climate risk and energy transition desks, agricultural economists, and data science portfolio reviewers.
The block above is the canonical text from docs/ABOUT.md, gated by a test. Edit it there.
TL;DR
- 🌍 ENSO Exposure Index. A world choropleth and leaderboard ranking where an ENSO swing reprices commodity and sector risk.
- 🔬 Causal rigor as the moat. Granger, CCM and a phase-randomized surrogate null separate real ENSO→price links from spurious ones. Not one of the seven tested links survives, and that is the honest headline.
- 📈 12-month forecasts. SARIMA plus LSTM ensemble, walk-forward backtested, beating persistence at all 12 leads.
- 🛰️ Live data. NOAA CPC ONI and the ENSO advisory fetched at runtime, ERSSTv5 SST grids, 71 World Bank commodities.
- 🇮🇳 Five region deep-dives. India (ENSO × Indian Ocean Dipole → monsoon → food CPI), SE Asia (palm oil), Brazil (coffee), Australia (wheat) and Peru (fishmeal), each ending in a computed desk view, including the ones where the honest view is no trade.
- 🔌 Machine-readable. Eight JSON endpoints under
/apiserve the same caches the pages read. - 🚀 Live and auto-deployed. Running on Hugging Face Spaces, CI/CD from GitHub.
The Macro Risk Desk landing: left rail (Niño-3.4 gauge, ONI trajectory, 12-month forecast cone), the world ENSO Exposure Index choropleth, a most-exposed-regions leaderboard, and the causation strip (the honesty layer):
|
🇮🇳 India deep-dive. ENSO × IOD → monsoon → food CPI, with a computed desk view and the real OLS regression heatmap. 
|
🛰️ Global SST Map. ERSSTv5 sea surface temperature anomalies with the classic El Niño equatorial Pacific warm tongue and its teleconnection zones. 
|
|
🔬 Causation Explorer. Live Granger and Convergent Cross Mapping on ONI against any commodity, with a plain-language verdict. 
|
📈 Forecast. SARIMA, LSTM and ensemble fan chart with confidence bands and ACC-vs-lead skill. 
|
Anyone can plot a correlation between El Niño and cocoa prices. The hard part, and the honest one, is asking does it survive a causal test? This desk runs two complementary engines on every ONI→commodity link:
- Granger causality (linear): does lagged ONI add predictive power over the commodity's own history?
- Convergent Cross Mapping / CCM (nonlinear, Sugihara et al. Science 2012): does cross-map skill rise and converge with library size in one direction only? Self-coded via simplex projection, with no
pyEDMdependency. - Phase-randomized surrogate significance (Ebisuzaki): does that skill beat a null built from the ONI's own power spectrum with the phases scrambled? This is the test that matters, because cross-map skill runs high between any two smooth seasonal series: two independent sine-plus-noise series score ρ ≈ 0.83 in this engine.
The result is deliberately humbling, and got more so. Of the seven ONI→commodity-price links tested, not one survives. All seven read weak, confounded. Palm oil and wheat read moderate until they were tested against the null (p = 0.15 and p = 0.47). The sharpest lesson is Robusta: it has the highest raw ρ on the board at 0.32 and the worst p-value at 0.976, against a null that averages ρ 0.23 on its own. That ρ 0.32 was previously quoted here as the desk's strongest causal evidence; it is indistinguishable from chance. The best link is Peru/fishmeal, at 21 of 24 Granger lags and ρ 0.29 against a 0.10 null, and even so, at p = 0.078, it misses the bar and stays weak. So the takeaway the desk leads with is:
Not one of the seven ENSO→commodity-price links the desk tests survives causal testing. The clean ENSO signal lives on the climate and production side, in the monsoon and Maritime Continent drought we prove in the region deep-dives, not in noisy monthly prices.
That "misattribution guard", showing the computed verdict instead of an asserted one even when it undercuts a tidy narrative, is the whole point. Palm oil and wheat both read moderate under the older raw-ρ rule and were quoted that way here; the surrogate null demoted both, and the desk now reports the demotion rather than the assumption that flattered them.
| # | Page | What it does |
|---|---|---|
| 00 | Macro Risk Desk (landing) | Command-bar terminal: Niño-3.4 gauge · ONI trajectory · forecast cone · ENSO Exposure Index choropleth · most-exposed leaderboard · causal-test strip |
| 01 | ENSO Monitor | Live ONI + RONI dual series (1950–present), gauge, live NOAA advisory badge, weekly Niño-3.4 nowcast (~1-week lag), CSV export |
| 02 | Global SST Map | ERSSTv5 2°×2° anomaly grids, flat + orthographic globe, teleconnection zones, optional EM-DAT disaster bubbles (±6 mo of the displayed field, sized by people affected) |
| 03 | Forecast | SARIMA + LSTM + ensemble fan chart, CI bands, ACC-vs-lead skill, observed-vs-forecast check (live weekly SST beside the ensemble's nearest month), analog panel, the nearest historical ENSO states by z-scored state vector, with their forward ONI paths |
| 04 | Sector Impact | Detrended lag-correlation heatmap, ONI × 71 commodities, lags 0–24 mo |
| 05 | Causation Explorer | Live Granger + CCM, both directions, plain-language verdict |
| 06 | Historical Events | Per-event cards since 1950: peak ONI/RONI, Callahan & Mankin 2023 GDP losses |
| 07 | 🇮🇳 India deep-dive | ENSO × IOD → monsoon → food CPI on IMD 0.25° gridded rainfall, area-weighted, 1901–2024; real OLS regression (n=124); desk view |
| 08 | 🌴 SE Asia deep-dive | Palm oil; the ENSO-premium story fails its own composite |
| 09 | 📐 Methodology | Renders docs/METHODOLOGY.md directly. One source of truth, no second copy to drift |
| 10 | 🩺 Source Status | Per-feed freshness, measured net of structural label lag, leading with "Behind" rather than "Age" |
| 11 | 🇧🇷 Brazil deep-dive | Arabica; the honest no-trade region. The phase composite hints at an El Niño premium, the lag profile finds nothing (peak r ≈ −0.07, on the window edge) |
| 12 | 🇦🇺 Australia deep-dive | Wheat; the drought is real and the price sign is inverted (r = −0.27 at 4 mo): correct physics, wrong instrument |
| 13 | 🇵🇪 Peru deep-dive | Fishmeal; the original El Niño and the desk's best near-miss, at 21/24 Granger lags and ρ 0.29 against a 0.10 null, surrogate p = 0.078 |
916 ENSO months · 42 events detected · 71 commodities · 2°×2° global SST grids from 1854 · 124 years of gridded Indian monsoon · 12-month forecast horizon · zero API keys required.
Positioning stances are computed, not typed: signed peak-lag correlation × current ENSO state, gated by the causal verdict, and haircut when the live observation diverges from the forecast. They move when the data moves, which is why this paragraph no longer quotes them. Because the surrogate null left all seven price links weak, the gate currently caps every region at WATCH; that is the honest reading, not a placeholder. Query /api/positioning for the live values.
Summary below. The full published spec, covering every weight, threshold, sample window, and the sources deliberately not used and why, is
docs/METHODOLOGY.md, served in-app as page 09 and checked against the code bytests/test_core.py.
Two models share an identical walk-forward verification harness, scored against a persistence reference (last observed ONI held constant):
| Model | Type | Architecture | Result |
|---|---|---|---|
| SARIMA | Statistical | statsmodels SARIMAX(2,0,1)(1,0,0,12) | Beats persistence at all 12 leads ✅ |
| LSTM | Deep Learning | PyTorch, 2-layer, 64 hidden | Beats persistence at all 12 leads ✅ |
Honest result: SARIMA outperforms the LSTM on this short univariate ONI series. The LSTM needs ancillary indices (IOD/MJO/PDO) or spatial SST fields (CNN track) to close the gap, and framing it otherwise would misrepresent the evidence. Both models' skill (ACC) drops below the 0.5 useful-skill threshold at 6–8 months, consistent with the ENSO spring predictability barrier.
Both tests run on linearly detrended (not differenced, because differencing kills the low-frequency ENSO band) ONI vs. commodity series:
-
Granger causality (linear): F-test across lags 0–24.
-
Convergent Cross Mapping (nonlinear): in-repo simplex projection (NumPy/SciPy), no pyEDM (its multiprocessing is incompatible with the Panel server on Windows). Genuine causation → cross-map skill rises and converges with library size in one direction only.
-
Surrogate significance (Ebisuzaki phase randomization, 500 draws per link): the observed ρ is scored against surrogates that keep the ONI's amplitude spectrum, meaning its annual cycle, persistence and smoothness, and randomize only the Fourier phases.
CAUSALandMODERATEnow additionally requirep < 0.05; a link that can't beat its own seasonal null is capped atWEAKwhatever its ρ, and an untested link (p = NaN) fails the gate rather than passing it.
The live explorer on page 05 does not run surrogates, because 500 extra cross-map passes per commodity is a precompute cost rather than a page-load cost, so its verdicts use the pre-surrogate rules and are exploratory. The landing strip carries the gated ones.
NOAA CPC · ERSSTv5 · World Bank · IMD data/ingest/ ──► data/cache/*.parquet
data/process/ ──► (phases · RONI · Granger+CCM · exposure index)
│
forecasting/ (SARIMA · LSTM · ensemble · skill) ──► forecasts/skill caches
│
▼
app.py ──► HoloViz Panel + Plotly ──► 00 Desk · 01 Monitor · 02 Map · 03 Forecast
04 Impact · 05 Causation · 06 History · 07 India · 08 SE Asia
09 Methodology · 10 Source Status · 11 Brazil
12 Australia · 13 Peru · 11 Brazil
12 Australia · 13 Peru · 11 Brazil
12 Australia · 13 Peru · 11 Brazil
12 Australia · 13 Peru · 11 Brazil
12 Australia · 13 Peru
│
Dockerfile ──► 🤗 Hugging Face Space (CI/CD from GitHub)
The dashboard reads parquet caches only. The heavy ingest and forecast pipeline (PyTorch, xarray, netCDF4) runs offline, so the deployed image is lean and serves instantly.
Every number on a page is also served as JSON, so a claim can be checked rather than screenshotted. Read-only, unauthenticated, CORS open, and backed by the same parquet caches the pages read.
| Endpoint | Returns |
|---|---|
/api |
Index of the endpoints below |
/api/state |
Current regime, phase, stance version, and the live weekly nowcast |
/api/positioning |
Computed stances per region, causal-gated |
/api/exposure |
ENSO Exposure Index per region and commodity |
/api/verdicts |
Granger and CCM verdicts per commodity, with surrogate p-values |
/api/analogs |
Nearest historical ENSO states and their forward ONI paths |
/api/sources |
Per-feed freshness, net of structural label lag |
/api/skill |
Forecast skill by lead |
/api/skill_variants |
Paired LSTM comparison, ONI-only against ONI plus 7 indices |
A missing cache returns 503 with a hint naming the script to run, never a 500. Serialization uses allow_nan=False, so a NaN fails loudly instead of emitting invalid JSON, and a test asserts it.
curl -s https://doginfantry-enso-macro-risk-desk.hf.space/api/verdictsEvery number on a page is also served as JSON, so a claim can be checked rather than screenshotted. Read-only, unauthenticated, CORS open, and backed by the same parquet caches the pages read.
| Endpoint | Returns |
|---|---|
/api |
Index of the endpoints below |
/api/state |
Current regime, phase, stance version, and the live weekly nowcast |
/api/positioning |
Computed stances per region, causal-gated |
/api/exposure |
ENSO Exposure Index per region and commodity |
/api/verdicts |
Granger and CCM verdicts per commodity, with surrogate p-values |
/api/analogs |
Nearest historical ENSO states and their forward ONI paths |
/api/sources |
Per-feed freshness, net of structural label lag |
/api/skill |
Forecast skill by lead |
/api/skill_variants |
Paired LSTM comparison, ONI-only against ONI plus 7 indices |
A missing cache returns 503 with a hint naming the script to run, never a 500. Serialization uses allow_nan=False, so a NaN fails loudly instead of emitting invalid JSON, and a test asserts it.
curl -s https://doginfantry-enso-macro-risk-desk.hf.space/api/verdictsEvery number on a page is also served as JSON, so a claim can be checked rather than screenshotted. Read-only, unauthenticated, CORS open, and backed by the same parquet caches the pages read.
| Endpoint | Returns |
|---|---|
/api |
Index of the endpoints below |
/api/state |
Current regime, phase, stance version, and the live weekly nowcast |
/api/positioning |
Computed stances per region, causal-gated |
/api/exposure |
ENSO Exposure Index per region and commodity |
/api/verdicts |
Granger and CCM verdicts per commodity, with surrogate p-values |
/api/analogs |
Nearest historical ENSO states and their forward ONI paths |
/api/sources |
Per-feed freshness, net of structural label lag |
/api/skill |
Forecast skill by lead |
/api/skill_variants |
Paired LSTM comparison, ONI-only against ONI plus 7 indices |
A missing cache returns 503 with a hint naming the script to run, never a 500. Serialization uses allow_nan=False, so a NaN fails loudly instead of emitting invalid JSON, and a test asserts it.
curl -s https://doginfantry-enso-macro-risk-desk.hf.space/api/verdictsEvery number on a page is also served as JSON, so a claim can be checked rather than screenshotted. Read-only, unauthenticated, CORS open, and backed by the same parquet caches the pages read.
| Endpoint | Returns |
|---|---|
/api |
Index of the endpoints below |
/api/state |
Current regime, phase, stance version, and the live weekly nowcast |
/api/positioning |
Computed stances per region, causal-gated |
/api/exposure |
ENSO Exposure Index per region and commodity |
/api/verdicts |
Granger and CCM verdicts per commodity, with surrogate p-values |
/api/analogs |
Nearest historical ENSO states and their forward ONI paths |
/api/sources |
Per-feed freshness, net of structural label lag |
/api/skill |
Forecast skill by lead |
/api/skill_variants |
Paired LSTM comparison, ONI-only against ONI plus 7 indices |
A missing cache returns 503 with a hint naming the script to run, never a 500. Serialization uses allow_nan=False, so a NaN fails loudly instead of emitting invalid JSON, and a test asserts it.
curl -s https://doginfantry-enso-macro-risk-desk.hf.space/api/verdictsEvery number on a page is also served as JSON, so a claim can be checked rather than screenshotted. Read-only, unauthenticated, CORS open, and backed by the same parquet caches the pages read.
| Endpoint | Returns |
|---|---|
/api |
Index of the endpoints below |
/api/state |
Current regime, phase, stance version, and the live weekly nowcast |
/api/positioning |
Computed stances per region, causal-gated |
/api/exposure |
ENSO Exposure Index per region and commodity |
/api/verdicts |
Granger and CCM verdicts per commodity, with surrogate p-values |
/api/analogs |
Nearest historical ENSO states and their forward ONI paths |
/api/sources |
Per-feed freshness, net of structural label lag |
/api/skill |
Forecast skill by lead |
/api/skill_variants |
Paired LSTM comparison, ONI-only against ONI plus 7 indices |
A missing cache returns 503 with a hint naming the script to run, never a 500. Serialization uses allow_nan=False, so a NaN fails loudly instead of emitting invalid JSON, and a test asserts it.
curl -s https://doginfantry-enso-macro-risk-desk.hf.space/api/verdictsRequires Python 3.12 (Hugging Face Spaces parity; the ML/geo stack lacks wheels on newer builds).
# 1. environment
py -3.12 -m venv .venv
.venv\Scripts\activate # Windows · source .venv/bin/activate on macOS/Linux
# 2. dependencies
pip install -r requirements.txt
# 3. (optional) refresh live data into data/cache/*.parquet
python data/ingest/oni_fetcher.py -v
python data/ingest/pink_sheet.py -v
# 4a. serve the whole site (landing at / plus all 13 sub-pages) via the unified entry point
python app.py # → http://localhost:5006
# 4b. or serve a single page
panel serve dashboard/pages/00_landing.py --showNo API keys required for any module. Every data source is free and public.
Live now: el-nino-green.vercel.app (Next.js front-door on Vercel, which prewarms the app host on visit) → huggingface.co/spaces/DogInfantry/enso-macro-risk-desk (the app itself, on the Docker SDK and free CPU Basic).
Panel/Bokeh is a long-running WebSocket server, so the app ships as a Docker Space (not Gradio/Static, and not Vercel without WASM conversion) running real panel serve via app.py. The web/ directory is a static-exported Next.js 15 landing page deployed on Vercel; it deep-links into the Space and polls the HF runtime API for live status. Data caches are refreshed monthly with one command: python scripts/refresh_data.py. A GitHub Action (.github/workflows/deploy-hf.yml) auto-syncs the Space on every push to master, so a push to GitHub makes the Space redeploy itself, and each run is recorded under the repo's Deployments tab. The serve-only dependency set (requirements-space.txt) excludes torch/xarray/kaleido, keeping the image small.
| Source | Provider | Module | Auth |
|---|---|---|---|
| ONI ASCII feed | NOAA CPC | oni_fetcher.py |
None |
| Weekly Niño-3.4 SST anomaly | NOAA CPC | weekly_nino34.py, read live at page load; the parquet snapshot is only the offline fallback |
None |
| ENSO Diagnostic Discussion (PDF) | NOAA CPC / IRI | advisory_fetcher.py |
None |
| Pink Sheet, monthly commodities | World Bank | pink_sheet.py |
None |
| ERSSTv5 netCDF grids | NOAA NCEI | ersst_fetcher.py |
None |
| IMD 0.25° daily gridded rainfall, 1901–2024 → cos(lat) area-weighted all-India JJAS | India Met. Dept. | imd_gridded.py, manual, ~3 GB of raw year-binaries; the parquet is committed |
None |
| IMD 36-subdivision monthly rainfall (1901–2017) | India Met. Dept. | monsoon_fetcher.py, superseded for all-India, kept for provenance |
None |
| SOI · Niño 1+2/3/4 · TNI · PDO · AMO · PNA · WP · DMI | NOAA PSL | climate_indices.py, where coverage() flags frozen upstreams, model_features() drops them before any model sees them |
None |
| EM-DAT disaster events (drought · flood · wildfire · storm) | CRED / UCLouvain | emdat_disasters.py, manual: neither the portal nor the HDX mirror permits automated download, so you export once by hand |
Free (registration) |
| ERA5 reanalysis · USDA NASS | Copernicus / USDA | (roadmap) | Free |
Rigorous analysis means disclosing limits. Read before drawing conclusions.
- ONI vs RONI. Charts label every index. On 16 Feb 2026 NOAA adopted RONI (subtracts tropical-mean SST to remove background warming) as the official ENSO index; under RONI the 2023–24 El Niño is ~0.6 °C cooler. Don't compare ONI- and RONI-classified events directly. This repo's RONI is computed from ERSSTv5 on a fixed 1991–2020 base, so it approximates the official value.
- The 3-month mean lags raw Niño-3.4, and is labelled by its centre month. A weekly spike can precede the smoothed ONI crossing ±0.5 °C by ~2 months. CPC's newest ONI row is a season: AMJ 2026 is stored under
2026-05-01, so a fully current reading legitimately displays as "May". Pages therefore name the season (AMJ 2026 · 3-mo mean, ctr. May) and pair it with the live weekly Niño-3.4 nowcast (~1-week lag) so freshness is visible. The weekly value is a different quantity, a single week of OISST rather than a 3-month mean, so it must never be read against ONI's ±0.5 °C event thresholds; the 4-week mean is shown beside it to damp noise. Current phase is fetched live, never hardcoded. - Correlation ≠ causation. Sector links are detrended Pearson r; the IOD and MJO can drive spurious co-movement. Causal direction needs Granger and CCM (Page 05), and every price link tested fails it (see The Moat).
- Exposure Index is a research construct: 50% computed peak lagged ONI–commodity correlation plus 50% curated structural exposure. Not an official product.
- Source freshness. The World Bank Pink Sheet workbook currently ends 2024-12; fetchers degrade gracefully to cache. India crop/CPI tabs are illustrative pending USDA/FAOSTAT ingestion. Page 10 (Source Status) publishes per-feed lateness live, measured net of each feed's structural label lag, because the ONI is stamped with its centre month and a perfectly current value therefore looks ~2.5 months old.
- The gridded all-India series approximates, not reproduces, IMD's published AISMR. The 1971–2020 normal comes out at 858.9 mm against IMD's ~868 (within 1.1%), and the departures track the old subdivision series at r=0.945, but IMD weights its own subdivisions rather than cos(lat) grid cells, so 2009 reads −15.0% here against a commonly cited ~−22%. Use these as this repo's series, never as IMD's official departures.
What is the ENSO Macro Risk Desk?
It's an interactive Python dashboard that maps the El Niño–Southern Oscillation (ENSO) cycle to commodity and sector risk, and stress-tests each link with causal inference (Granger + Convergent Cross Mapping). It's built for commodity/macro analysts and climate-risk desks, and it's live on Hugging Face Spaces.
What is ENSO, and why does it matter for commodity markets?
ENSO is the dominant year-to-year driver of global climate variability. El Niño (ONI ≥ +0.5 °C) suppresses rainfall in Southeast Asia and Australia and enhances it on South America's west coast; La Niña (ONI ≤ −0.5 °C) reverses it. Because ENSO disrupts rainfall in the world's key agricultural zones, it is linked to price shocks in wheat, maize, rice, coffee, cocoa, and palm oil, typically 3–9 months after the SST anomaly peak.
What is the difference between ONI and RONI?
ONI is the 3-month running mean of Niño-3.4 SST anomalies against a rolling 30-year base. RONI (Relative ONI) subtracts the tropical-mean (20°S–20°N) anomaly, removing the warming trend that increasingly inflates ONI. NOAA adopted RONI as the official ENSO index in February 2026; under it the 2023–24 El Niño is ~0.6 °C weaker.
What is Convergent Cross Mapping (CCM)?
CCM (Sugihara et al., Science, 2012) is a nonlinear causal-inference method for dynamical systems. It tests whether X drives Y by checking whether Y's reconstructed attractor can recover X's states, and whether that cross-map skill converges as the observation library grows. Unlike Granger causality it doesn't assume linearity. This project implements CCM in-repo via simplex projection (NumPy/SciPy), without pyEDM, for Windows/Panel compatibility.
Is it really live? How is it deployed?
Yes, running on Hugging Face Spaces as a Docker Space serving panel serve via app.py, on the free CPU tier (it may cold-start after inactivity). A GitHub Action auto-redeploys on every push to master.
Why does SARIMA outperform the LSTM?
ONI is a short (~70-year) quasi-periodic univariate signal that SARIMA exploits directly through its seasonal AR structure. Without ancillary indices (IOD/MJO/PDO) or spatial SST input (CNN track), the LSTM lacks the signal to overcome SARIMA's parsimony at this data scale. That is an honest and common finding for short univariate climate series.
Can I run it on macOS or Linux?
Yes. Swap .venv\Scripts\activate for source .venv/bin/activate; the rest is cross-platform. The only Windows-specific choice is avoiding pyEDM, and the in-repo CCM is fully portable.
Data © respective providers: NOAA/NWS (ONI, advisory, ERSSTv5), World Bank (Pink Sheet), India Meteorological Department (monsoon), Copernicus/ECMWF, USDA, FAO, CRED (EM-DAT). Code and dashboard: research and educational use. Cite the primary data sources when reusing outputs.
Key reference: Callahan, C. W. & Mankin, J. S. (2023). Persistent effect of El Niño on global economic growth. Science, 381, 789–793. DOI: 10.1126/science.adf0374
Built with HoloViz Panel, Plotly, PyTorch, statsmodels, and data from NOAA CPC and the World Bank.
Topics: el-nino · la-nina · enso · enso-forecast · climate-risk · commodity-markets · macro-research · teleconnections · oceanic-nino-index · oni · roni · nino-3.4 · granger-causality · convergent-cross-mapping · ccm · sarima · lstm · time-series-forecasting · climate-finance · indian-ocean-dipole · monsoon · palm-oil · sea-surface-temperature · python-dashboard · holoviz-panel · plotly · data-visualization