An analysis of animal-related requests completed by NYC Park Rangers since 2018, animal characteristics, locations, response times, and temporal patterns. Generated by Mistral Vibe Chat in Think mode.

NYC Park Ranger Animal Request Trends

See what tables I have in the database my_shared_db.

Describe the schema of the `nyc_park_ranger_completed_animal_request_data` table, what the columns in the table contain, and how the columns are connected.

Create a Dive that illustrates trends in the requests in the table pertaining to the animals (their species, condition, and age, and how many are present) that trigger a request, where they're found (borough, property, and location), how long it takes for the ranger to respond to a request after the request is first logged, and how these metrics have changed with season, day of the week, and time of the day. Take your time to think carefully about the cleanest, most robust way to do this before doing it.

I think this is great! Here is my feedback:
* Let's use a pie chart to illustrate the full species distribution in the data
* Let's also show trends in requests and avg response time by borough by hour of the day, day of the week, and the month when the request was first logged, if possible
* Let's also show trends in animal species by borough and most common property, if possible

On 2nd thought, please revert from the pie chart illustrating the full species distribution in the data to a bar chart of the 10 most common species (your original design for this metric). And if you think this would be good for visibility, please change the species by location and property charts to stacked bar charts showing the top 3-5 most common species by location or property.

This is great! Some charts are missing some tick labels for their x- or y-axis. Please resize the chart, change the tick label text size, and/or rearrange the layout so that no chart is missing tick labels, making sure that the final results are as clean, clear, and easy to read as possible.

1 last request: could you resize the scrollable legend for the Top 3 Species by Property (Stacked) chart and/or move it to the blank portion of the chart so that it doesn't obscure the tick labels? Also, please make sure that all the tick labels for the chart are present and visible.

Sorry, but 1 last last request: do you think it would be cleaner to change the Top 3 Species by Property (Stacked) chart to a horizontal stacked bar chart? Currently the tick labels are being cut off. Perhaps it would help visibility to have the Top 3 Species by Borough (Stacked) chart occupy the top half of the page, then the Top 3 Species by Property (Stacked) chart as a horizontal stacked bar chart in the bottom half of the page with its legend.

Amy Ngo


import { useState, useMemo } from "react";
import { useSQLQuery } from "@motherduck/react-sql-query";
import { BarChart, Bar, LineChart, Line, XAxis, YAxis, CartesianGrid, Tooltip, ResponsiveContainer, ReferenceLine, Cell } from "recharts";

const N = (v) => (v != null ? Number(v) : 0);
const WHO = 5;
const CSV = `read_csv_auto('https://us.data.motherduck.com/who_ambient_air_quality/csv/who_ambient_air_quality_database_version_2024.csv')`;

const CITIES_LIST = [
  "Addis Ababa/ETH","Aix En Provence/FRA","Ankara/TUR","Annecy/FRA","Antwerpen/BEL",
  "Athina/GRC","Bangkok/THA","Barcelona/ESP","Basel/CHE","Beijing/CHN",
  "Belgrade/SRB","Bergamo/ITA","Berlin/DEU","Bern/CHE","Bogota/COL",
  "Bologna/ITA","Bordeaux/FRA","Boston Cambridge Newton MA NH/USA","Bratislava/SVK",
  "Brescia/ITA","Brno/CZE","Brussels/BEL","Budapest/HUN",
  "Chicago Naperville Elgin IL IN WI/USA","Clermont Ferrand/FRA","Colombo/LKA",
  "Delhi/IND","Denver Aurora Lakewood CO/USA","Dhaka/BGD","Dijon/FRA","Dubai/ARE",
  "Dublin/IRL","Gent/BEL","Grenoble/FRA","Hamburg/DEU","Hanoi/VNM",
  "Helsinki Helsingfors/FIN","Istanbul/TUR","Jakarta/IDN","Johannesburg/ZAF",
  "Kathmandu/NPL","Krakow/POL","Lausanne/CHE","Liege/BEL","Lille/FRA","Lima/PER",
  "Ljubljana/SVN","London/GBR","Los Angeles Long Beach Anaheim CA/USA","Lyon/FRA",
  "Madrid/ESP","Marseille/FRA","Montpellier/FRA","Montreal/CAN","Mumbai/IND",
  "Nairobi/KEN","Namur/BEL","Nantes/FRA","Nice/FRA","Oslo/NOR","Paris/FRA",
  "Rennes/FRA","Riga/LVA","San Francisco Oakland Hayward CA/USA","Santiago/CHL",
  "Sao Paulo/BRA","Seattle Tacoma Bellevue WA/USA","Seoul/KOR","Shanghai/CHN",
  "Singapore/SGP","Stockholm/SWE","Strasbourg/FRA","Sydney/AUS","Tallinn/EST",
  "Tokyo/JPN","Toronto/CAN","Toulouse/FRA","Vancouver/CAN","Vilnius/LTU",
  "Zagreb/HRV","Zurich/CHE"
];

const RL = {"1_Afr":"Africa","2_Amr":"Americas","3_Sear":"South-East Asia","4_Eur":"Europe","5_Emr":"Eastern Mediterranean","6_Wpr":"Western Pacific"};

function tier(v) {
  if (v <= 5) return { l: "Good", c: "#2d7a00" };
  if (v <= 15) return { l: "Moderate", c: "#e18727" };
  if (v <= 35) return { l: "High", c: "#bd4e35" };
  return { l: "Hazardous", c: "#bc1200" };
}

const CT = ({ active, payload }) => {
  if (!active || !payload?.[0]) return null;
  const d = payload[0].payload;
  return (
    <div style={{ background: "#231f20", color: "#fff", padding: "6px 10px", borderRadius: 4, fontSize: 11, fontFamily: "monospace" }}>
      <div style={{ fontWeight: 600 }}>{d.city || d.year}</div>
      <div>PM2.5: {N(d.pm25).toFixed(1)} μg/m³</div>
    </div>
  );
};

export default function WHOAirQuality() {
  const [sel, setSel] = useState("Paris/FRA");
  const [yr, setYr] = useState(null);

  const safeSel = sel.replace(/'/g, "''");

  const trend = useSQLQuery(`
    SELECT TRY_CAST(year AS INT) as year,
           ROUND(TRY_CAST(pm25_concentration AS DOUBLE), 1) as pm25,
           who_region, country_name
    FROM ${CSV}
    WHERE city = '${safeSel}'
      AND TRY_CAST(pm25_concentration AS DOUBLE) IS NOT NULL
      AND TRY_CAST(year AS INT) IS NOT NULL
    ORDER BY TRY_CAST(year AS INT)
  `);

  const trendRows = Array.isArray(trend.data) ? trend.data : [];
  const years = trendRows.map(r => N(r.year));
  const maxYear = years.length > 0 ? Math.max(...years) : null;
  const effectiveYr = yr !== null && years.includes(yr) ? yr : maxYear;
  const region = trendRows[0]?.who_region || "";
  const country = trendRows[0]?.country_name || "";

  const curRow = trendRows.find(r => N(r.year) === effectiveYr);
  const pm = curRow ? N(curRow.pm25) : 0;
  const t = tier(pm);

  const peers = useSQLQuery(`
    SELECT city, ROUND(TRY_CAST(pm25_concentration AS DOUBLE), 1) as pm25
    FROM ${CSV}
    WHERE who_region = '${region}'
      AND TRY_CAST(year AS INT) = ${effectiveYr}
      AND TRY_CAST(pm25_concentration AS DOUBLE) IS NOT NULL
    ORDER BY TRY_CAST(pm25_concentration AS DOUBLE) ASC
    LIMIT 150
  `, { enabled: !!region && effectiveYr !== null });

  const peerRows = Array.isArray(peers.data) ? peers.data : [];
  const rankIdx = peerRows.findIndex(r => r.city === sel);
  const rank = rankIdx >= 0 ? rankIdx + 1 : null;

  const barData = useMemo(() => {
    if (!peerRows.length) return [];
    const idx = peerRows.findIndex(r => r.city === sel);
    const center = idx >= 0 ? idx : 0;
    const s = Math.max(0, center - 8);
    return peerRows.slice(s, s + 18).map(r => ({ city: String(r.city), pm25: N(r.pm25) }));
  }, [peerRows, sel]);

  const trendData = trendRows.map(r => ({ year: N(r.year), pm25: N(r.pm25) }));
  const cityName = sel.split("/")[0];

  return (
    <div className="p-4" style={{ background: "#f8f8f8", color: "#231f20" }}>
      <h1 className="text-2xl font-semibold">WHO Air Quality Index</h1>
      <p className="text-sm mb-4" style={{ color: "#6a6a6a" }}>PM2.5 Annual Mean · WHO Database v2024</p>

      <div className="flex items-end gap-4 mb-5 flex-wrap">
        <div>
          <p className="text-xs mb-1 font-medium" style={{ color: "#6a6a6a", textTransform: "uppercase", letterSpacing: "1px" }}>City</p>
          <select value={sel} onChange={e => { setSel(e.target.value); setYr(null); }}
            className="text-sm p-2 rounded"
            style={{ border: "1px solid #ddd", background: "#fff", minWidth: 220 }}>
            {CITIES_LIST.map(c => <option key={c} value={c}>{c}</option>)}
          </select>
        </div>
        <div>
          <p className="text-xs mb-1 font-medium" style={{ color: "#6a6a6a", textTransform: "uppercase", letterSpacing: "1px" }}>Year</p>
          <div className="flex gap-1 flex-wrap">
            {trend.isLoading ? (
              <div className="h-7 w-32 bg-gray-200 animate-pulse rounded" />
            ) : years.map(y => (
              <button key={y} onClick={() => setYr(y)}
                className="text-xs px-2 py-1 rounded"
                style={{
                  background: y === effectiveYr ? "#231f20" : "#fff",
                  color: y === effectiveYr ? "#fff" : "#231f20",
                  border: "1px solid #ddd", fontFamily: "monospace", cursor: "pointer"
                }}>
                {y}
              </button>
            ))}
          </div>
        </div>
      </div>

      <div className="grid grid-cols-3 gap-8 mb-5">
        <div>
          {trend.isLoading ? <div className="h-12 w-24 bg-gray-200 animate-pulse rounded" /> : (
            <p className="text-5xl font-bold" style={{ color: t.c, fontFamily: "monospace" }}>{pm.toFixed(1)}</p>
          )}
          <p className="text-sm mt-1" style={{ color: "#6a6a6a" }}>PM2.5 μg/m³ · {trend.isLoading ? "…" : t.l}</p>
        </div>
        <div>
          {trend.isLoading ? <div className="h-12 w-24 bg-gray-200 animate-pulse rounded" /> : (
            <p className="text-5xl font-bold" style={{ color: pm <= 5 ? "#2d7a00" : "#bc1200", fontFamily: "monospace" }}>
              {pm <= 5 ? "✓" : `${(pm / WHO).toFixed(1)}×`}
            </p>
          )}
          <p className="text-sm mt-1" style={{ color: "#6a6a6a" }}>vs WHO guideline (5 μg/m³)</p>
        </div>
        <div>
          {peers.isLoading ? <div className="h-12 w-24 bg-gray-200 animate-pulse rounded" /> : (
            <p className="text-5xl font-bold" style={{ color: "#231f20", fontFamily: "monospace" }}>
              {rank ? `#${rank}` : "—"}
            </p>
          )}
          <p className="text-sm mt-1" style={{ color: "#6a6a6a" }}>
            {rank ? `of ${peerRows.length} in ${RL[region] || "region"}` : RL[region] || ""}
          </p>
        </div>
      </div>

      <p className="text-xs font-medium mb-2" style={{ color: "#6a6a6a", textTransform: "uppercase", letterSpacing: "1px" }}>
        Regional Peers · {RL[region] || ""} · {effectiveYr || "—"}
      </p>
      {peers.isLoading ? (
        <div className="bg-gray-100 animate-pulse rounded" style={{ height: 200 }} />
      ) : barData.length > 0 ? (
        <ResponsiveContainer width="100%" height={200}>
          <BarChart data={barData} margin={{ top: 4, right: 8, bottom: 44, left: 4 }}>
            <CartesianGrid strokeDasharray="3 3" stroke="#eee" vertical={false} />
            <XAxis dataKey="city" tick={{ fontSize: 8, fill: "#6a6a6a" }} angle={-45} textAnchor="end" height={60} interval={0}
              tickFormatter={v => v.length > 12 ? v.slice(0, 10) + "…" : v} />
            <YAxis tick={{ fontSize: 10, fill: "#6a6a6a" }} />
            <Tooltip content={<CT />} />
            <ReferenceLine y={WHO} stroke="#2d7a00" strokeDasharray="4 3" strokeWidth={1.5} />
            <Bar dataKey="pm25" radius={[2, 2, 0, 0]} maxBarSize={18}>
              {barData.map((d, i) => (
                <Cell key={i} fill={d.city === sel ? "#0777b3" : "#adadad"} opacity={d.city === sel ? 1 : 0.45} />
              ))}
            </Bar>
          </BarChart>
        </ResponsiveContainer>
      ) : (
        <div style={{ height: 200, display: "flex", alignItems: "center", justifyContent: "center", color: "#6a6a6a", fontSize: 13 }}>
          No peer data available
        </div>
      )}

      <p className="text-xs font-medium mt-4 mb-2" style={{ color: "#6a6a6a", textTransform: "uppercase", letterSpacing: "1px" }}>
        PM2.5 Trend · {cityName}, {country}
      </p>
      {trend.isLoading ? (
        <div className="bg-gray-100 animate-pulse rounded" style={{ height: 200 }} />
      ) : trendData.length > 1 ? (
        <ResponsiveContainer width="100%" height={200}>
          <LineChart data={trendData} margin={{ top: 8, right: 24, bottom: 4, left: 4 }}>
            <CartesianGrid strokeDasharray="3 3" stroke="#eee" vertical={false} />
            <XAxis dataKey="year" tick={{ fontSize: 11, fill: "#6a6a6a" }} />
            <YAxis tick={{ fontSize: 11, fill: "#6a6a6a" }} domain={[0, "auto"]} />
            <Tooltip content={<CT />} />
            <ReferenceLine y={WHO} stroke="#2d7a00" strokeDasharray="4 3" strokeWidth={1.5} />
            <Line type="linear" dataKey="pm25" stroke="#0777b3" strokeWidth={2} dot={{ r: 3, fill: "#0777b3" }} />
          </LineChart>
        </ResponsiveContainer>
      ) : (
        <div style={{ height: 200, display: "flex", alignItems: "center", justifyContent: "center", color: "#6a6a6a", fontSize: 13 }}>
          {trendData.length === 1 ? "Single data point — trend requires multiple years" : "No data available"}
        </div>
      )}

      <p className="text-xs mt-3 text-center" style={{ color: "#6a6a6a" }}>
        WHO Ambient Air Quality Database v2024 · Queried live via DuckDB on MotherDuck
      </p>
    </div>
  );
}


Interactive PM2.5 explorer — compare cities, track trends, and benchmark against WHO guidelines. Data from WHO Ambient Air Quality Database v2024, queried live via DuckDB.

WHO Air Quality Index

Mehdi Ouazza

IRC impact explorer governed by the canonical 2026-09-15 source-of-truth workbook and external methods note. Our Reach shows FY25 reported service outcomes; Where Next shows where additional IRC funding could stretch furthest using modeled average IRC delivery cost at significant scale; What's Possible uses funding-first planning estimates of annual delivery at significant scale. Public provenance is deliberately concise, with exhaustive assumptions and release controls maintained in the canonical workbook.

Where can IRC have the greatest impact?

Meridian — Where can IRC have the greatest impact?

Andrew Wilcox

Get Control Tower at: https://github.com/SDB-Ryan/motherduck-control-tower.git

Control Tower is a self-updating map of a MotherDuck account: every dive, flight, table, and share, and how data flows between them, with live row counts and run health on every node. 

You never draw the graph — you catalog each object's lineage once, and a scheduled flight builds the picture. Stale tables and failed runs light up on the board; anything uncataloged lands on a warnings strip instead of going missing. 

Free and open source — hand the repo's INSTALL.md to your AI assistant and it installs it on your own account.

Control Tower

Control Tower - Observability for databases, flights and dives

Ryan Dolley

Google Analytics-style dashboard for Dive usage in your org, built entirely on MD_INFORMATION_SCHEMA.QUERY_HISTORY (user_agent dive IDs) + MD_LIST_DIVES(). GA4 Material visual language: white cards, scorecards with delta arrows, blue/gray palette, date range top-right. Snapshot with PoP deltas, org totals incl. endorsed dive count, trend, hour/weekday heatmap. Dives tab with sparklines, status pills, Δ columns, database breakdown, needs-attention (declining + unused + endorsed coverage). Audience with new vs returning, per-user Δ, surface split, creator leaderboard with endorsed counts. Performance with p95/avg exec, wait time, spill. Clickable drill-downs to dive/user details incl. error breakdown. CSV exports (disabled without access). Graceful no-permission handling: single gated probe, warning banner, empty states instead of errors. Requires Business plan + org admin.

Dive Analytics

Luka Lovosevic

An interactive cinematic journey through the DuckDB ecosystem — from a single research project to an entire galaxy of tools. Hover, click, and explore the universe.

Duckverse Chronicles

Rishabh Sethi

Hacker News stories and comments naming OpenAI, Anthropic, Google Gemini or DeepSeek (2019 to Sept 2026): 488k regex matches, a 65,635-item stratified sample classified with prompt_jev() (refers, relevant, negative, topic in one call) and weighted back. Share of the conversation per quarter, negativity per half-year, complaint topics per lab with example comments, and the name check (Claude Shannon, Gemini protocol...).

What does Hacker News think of each AI lab?

Live Bern 10-minute temperature vs the fitted heating threshold, daily consumption-vs-temperature chart with the hinge fit, and the consumption-weighted station map. Refreshed by Flight swiss-heating-threshold.

Trading desk: weather and Swiss power demand

I'd like to **understand how strongly temperature drives Swiss electricity consumption this year**.
I have two data sources. Swissgrid publishes the "Energieübersicht Schweiz" as an Excel file every month, you'll find it here: https://www.swissgrid.ch/dam/jcr:805e525c-44fe-4701-a227-6144193257ac/EnergieUebersichtCH-2026.xlsx. The sheet "Zeitreihen0h15" contains 15-minute values in kWh. The column I care about is "Summe endverbrauchte Energie Regelblock Schweiz" (end-user consumption), not total consumption, because the total includes pump storage and grid losses. The sheet also has consumption per canton further to the right, but note that Swissgrid groups some cantons together, for example BE with JU and ZH with SH.

For the weather, MeteoSwiss publishes all its automatic stations as open data here: https://data.geo.admin.ch/browser/index.html#/collections/ch.meteoschweiz.ogd-smn. Every station has a folder with daily, hourly and 10-minute files. The daily files are what I need. The parameter for daily mean air temperature at 2 metres is called tre200d0. There's a station list with coordinates and canton in ogd-smn_meta_stations.csv.

Please start with Bern (station BER) as a proxy for the Swiss Plateau. Give me the daily end-user consumption against the daily mean temperature for 2026 so far, and tell me from which temperature downwards consumption starts rising. In our industry we call that the heating threshold, we usually assume around 12 °C. I want to see if that still holds. 

Also tell me if anything in the data looks off before you trust it.

You suggested a weighted temperature across stations instead of Bern only. Agreed. Use one representative station per canton, weight each canton by its share of consumption in the Swissgrid canton columns, and skip stations above 1500 m, nobody heats a home up there. Then show me the stations on a map of Switzerland with the canton outlines, coloured by this year's mean temperature. Canton outlines you can take from here: https://raw.githubusercontent.com/codeforgermany/click_that_hood/main/public/data/switzerland.geojson

Note: the outlines are rough, trust the canton column in the station list.

What's Bern doing right now, has it dropped below 13 °C in the last 24 hours? I want the 10-minute values, not daily averages.

Then put everything into one dashboard for the trading desk: the threshold chart, the map, and the live Bern temperature at the top. Add a button that pulls fresh data, and make it update itself every morning at 7 before the desk opens.

Simon

A medieval-market interactive explainer for Old School RuneScape’s Grand Exchange, using the full live item universe plus focused ranked ledgers to teach spread, tax, liquidity, Bonds, market depth, and purchasing power.

Dive Gallery