30:T6b7,# What's it like on Lake Michigan on the North Side of Chicago right now?

This notebook queries a sensor platform at the [Chicago Pier Buoy](https://seagull.glos.org/data-console/98) on Lake Michigan north of downtown Chicago using the [Seagull API](https://seagull.glos.org/data-console). I like to go running on Lake Michigan, so I made a bit of code that writes simple sentences that help me know what the swimming conditions will be like. Is it warm enough to take a dip after my run? Are the waves so big that if the mean lifeguard is on duty, she will chase me off, and I should wait until after hours? In the summer when I lived in Chicago, I checked this every day! This is an example about how to use the tools of data journalism to independently interrogate your world. I did research to understand things like historic surface temperatures and what medical professionals consider dangerous. I combined objective analysis with subjective experience by using a statistical technique called z-scores to make it easier for me to map my experience, perception, and judgement onto the data. You'll see on a day with big waves that the tool says you might encounter surfers -- that's because I saw people surfing on days when the score exceeded a certain threshold. 31:T69e,waveChart = { const maxWidth = width > 750 ? 750 : width; return Plot.plot({ width: maxWidth, height: maxWidth * 0.7, style: { fontSize: "12px" }, marginRight: 125, marginLeft: 10, marginTop: 30, marginBottom: 40, color: { scheme: "Magma", domain: [10, 26] }, x: { ticks: 4, clamp: true, domain: [chartRangeStart, date_range.get("max_date", 0)] }, y: { label: `Surface temp (c)`, nice: true, ticks: false, clamp: true, domain: [10, 26] }, marks: [ Plot.rect(chartZones, { x1: date_range.get("min_date", 0), x2: date_range.get("max_date", 0), y1: summary .filter((d) => aq.op.equal(d.parameter_label, "Surface temp (degrees Celsius)") ) .get("min", 0), y1: 10, y2: "temp", fill: (d, i) => { return schemeBlues[i]; } }), Plot.text(chartZones, { x: date_range.get("max_date", 0), y: "temp", text: "label", lineAnchor: "top", textAnchor: "start", fontWeight: "bold", fontSize: (d) => (d.temp === 22 ? "18px" : "18px"), dy: 5, dx: 5 }), Plot.text(chartZones, { x: date_range.get("max_date", 0), y: "temp", text: "description", lineAnchor: "top", textAnchor: "start", fontSize: (d) => "14px", dy: 22, dx: 5 }), Plot.lineY(temperatureChartData, { x: "date", y: "rolling_average", strokeWidth: 4, z: null, stroke: "rolling_average" }) ] }); }32:T920,

Near my beach right now...

${getThresholdLabel( recent .filter((d) => d.parameter_label === "Significant wave height (meters)") .get("last_z", 0), wave_height_thresholds )} (now: ${recent .filter((d) => d.parameter_label === "Significant wave height (meters)") .get("last_value", 0).toLocaleString(undefined, {maximumFractionDigits: 2})}m, mean: ${summary .filter((d) => d.parameter_label === "Significant wave height (meters)") .get("mean", 0).toLocaleString(undefined, {maximumFractionDigits: 2})}m, z-score: ${recent .filter((d) => d.parameter_label === "Significant wave height (meters)") .get("last_z", 0).toLocaleString(undefined, {maximumFractionDigits: 1})}).
${getThresholdLabel( recent .filter((d) => d.parameter_label === "Mean wave period (seconds)") .get("last_z", 0), wave_period_thresholds )} (now: a wave every ${recent .filter((d) => d.parameter_label === "Mean wave period (seconds)") .get("last_value", 0).toLocaleString(undefined, {maximumFractionDigits: 1})} seconds, mean: a wave every ${summary .filter((d) => d.parameter_label === "Mean wave period (seconds)") .get("mean", 0).toLocaleString(undefined, {maximumFractionDigits: 1})} seconds, z-score: ${recent .filter((d) => d.parameter_label === "Mean wave period (seconds)") .get("last_z", 0).toLocaleString(undefined, {maximumFractionDigits: 1})}).
${getThresholdLabel( recent .filter((d) => d.parameter_label === "Surface temp (degrees Celsius)") .get("last_value", 0), temp_thresholds )} (now: ${recent .filter((d) => d.parameter_label === "Surface temp (degrees Celsius)") .get("last_value", 0).toLocaleString(undefined, {maximumFractionDigits: 1})}C, mean: ${summary .filter((d) => d.parameter_label === "Surface temp (degrees Celsius)") .get("mean", 0).toLocaleString(undefined, {maximumFractionDigits: 1})}C).

The most recent measurement was ${date_range.get("max_date", 0)}}. All historical measurements, like mean surface temperature, are based on data going back to ${date_range.get("min_date", 0)}. Qualitative assessments, like whether the water is chilly or waves are small, reflect my own observations.

33:Te42,# How does it work? * Get as many historical measurements as possible up to the most recent available from the Seagull API. I had to figure out which sensor I wanted to query by visually inspecting available sensors near where I like to run. * Did a bit of research into which measurements might be most significant, and also went to the lake to see which ones corresponded to my perceptions. * Assign a z-score to each measurement across the available historical data: * -3/+3 are abnormally high or low. * -1/+1 are more typical. * Assign a qualitative label to the most recent measurement: * For wave speed and size, use z-score to compare to recent history. A wave height with a z-score of +3 is quite large compared to past 1-3 months. * For temperature, use the absolute value. Under 0C means the surface is frozen, 0-8C is a no-go for me, 8-12C isn't particularly comfortable, etc. In practical terms, this means I wrote some code that maps words to values. ## Wave size and speed Using z-score across as little as one recent month of data has been more than adequate at conveying useful wave information, in spite of my fears that data would be required from multiple years to understand what's normal and what's an outlier. My regular trips to the lake indicated this was an unfounded concern: If the wave height had a z-score of -3, the lake was like a mirror, and if it was +3, there were white caps and big slow waves coming in. This has led me to surmise that waves on this part of Lake Michigan don't tend to exhibit extreme behavior. z-score also works well because it's very simple to remember when I'm out "observing" and don't want to be thinking at all about technology. My lizard brain, even when occupied with my aching knees, is still pretty OK with 3 = large size/low frequency, 1 = normal, -3 = small size/high frequency. As a sanity check, I looked into typical summer wave heights and compared to what I was seeing. A [2014 Chicago Tribune article](https://web.archive.org/web/20210624085218/https://www.chicagotribune.com/weather/ct-wea-0817-asktom-20140817-column.html) says summer waves are usually 2 to 4 feet, though it does not specify the type of wave height. So far in 2023, the median significant wave height has been around 0.5m, or somewhat less than 2 feet. This isn't strong agreement, nor is it is unreasonably far off. The lookup for wave height uses z-scores: ```js wave_height_thresholds = [ { min: 3, max: null, label: "The waves might be tall enough to surf on, the beach is definitely closed" }, { min: 1, max: 3, label: "The waves are big and the beach might be closed" }, # ... { min: null, max: -3, label: "The lake is a mirror" } ] ``` The opposite is true of temperature. ## Temperature Using a z-score derived from limited historical data works poorly when it comes to temperature. In the early spring, 12C might be quite warm compared to when there were ice chunks floating in the lake a month ago. That's not what we're shooting for. Instead, we want to use an absolute scale. The lookup uses absolute temperature ranges: ```js temp_thresholds = [ { min: null, max: 0, label: "The temperature means the surface is frozen" }, # ... { min: 20, max: 23, label: "The temperature is at the summer peak, approaching the temperature of a pool" } ] ``` The historical data used to go back 3 months, which appeared to be about the point where the API would stop sending data. However, the sensor went down during the winter of 2022-2023, and only came back in late April, 2023. 34:Td10,lotsOfBlues = [ "#f7fbff", "#f6faff", "#f5fafe", "#f5f9fe", "#f4f9fe", "#f3f8fe", "#f2f8fd", "#f2f7fd", "#f1f7fd", "#f0f6fd", "#eff6fc", "#eef5fc", "#eef5fc", "#edf4fc", "#ecf4fb", "#ebf3fb", "#eaf3fb", "#eaf2fb", "#e9f2fa", "#e8f1fa", "#e7f1fa", "#e7f0fa", "#e6f0f9", "#e5eff9", "#e4eff9", "#e3eef9", "#e3eef8", "#e2edf8", "#e1edf8", "#e0ecf8", "#e0ecf7", "#dfebf7", "#deebf7", "#ddeaf7", "#ddeaf6", "#dce9f6", "#dbe9f6", "#dae8f6", "#d9e8f5", "#d9e7f5", "#d8e7f5", "#d7e6f5", "#d6e6f4", "#d6e5f4", "#d5e5f4", "#d4e4f4", "#d3e4f3", "#d2e3f3", "#d2e3f3", "#d1e2f3", "#d0e2f2", "#cfe1f2", "#cee1f2", "#cde0f1", "#cce0f1", "#ccdff1", "#cbdff1", "#cadef0", "#c9def0", "#c8ddf0", "#c7ddef", "#c6dcef", "#c5dcef", "#c4dbee", "#c3dbee", "#c2daee", "#c1daed", "#c0d9ed", "#bfd9ec", "#bed8ec", "#bdd8ec", "#bcd7eb", "#bbd7eb", "#b9d6eb", "#b8d5ea", "#b7d5ea", "#b6d4e9", "#b5d4e9", "#b4d3e9", "#b2d3e8", "#b1d2e8", "#b0d1e7", "#afd1e7", "#add0e7", "#acd0e6", "#abcfe6", "#a9cfe5", "#a8cee5", "#a7cde5", "#a5cde4", "#a4cce4", "#a3cbe3", "#a1cbe3", "#a0cae3", "#9ec9e2", "#9dc9e2", "#9cc8e1", "#9ac7e1", "#99c6e1", "#97c6e0", "#96c5e0", "#94c4df", "#93c3df", "#91c3df", "#90c2de", "#8ec1de", "#8dc0de", "#8bc0dd", "#8abfdd", "#88bedc", "#87bddc", "#85bcdc", "#84bbdb", "#82bbdb", "#81badb", "#7fb9da", "#7eb8da", "#7cb7d9", "#7bb6d9", "#79b5d9", "#78b5d8", "#76b4d8", "#75b3d7", "#73b2d7", "#72b1d7", "#70b0d6", "#6fafd6", "#6daed5", "#6caed5", "#6badd5", "#69acd4", "#68abd4", "#66aad3", "#65a9d3", "#63a8d2", "#62a7d2", "#61a7d1", "#5fa6d1", "#5ea5d0", "#5da4d0", "#5ba3d0", "#5aa2cf", "#59a1cf", "#57a0ce", "#569fce", "#559ecd", "#549ecd", "#529dcc", "#519ccc", "#509bcb", "#4f9acb", "#4d99ca", "#4c98ca", "#4b97c9", "#4a96c9", "#4895c8", "#4794c8", "#4693c7", "#4592c7", "#4492c6", "#4391c6", "#4190c5", "#408fc4", "#3f8ec4", "#3e8dc3", "#3d8cc3", "#3c8bc2", "#3b8ac2", "#3a89c1", "#3988c1", "#3787c0", "#3686c0", "#3585bf", "#3484bf", "#3383be", "#3282bd", "#3181bd", "#3080bc", "#2f7fbc", "#2e7ebb", "#2d7dbb", "#2c7cba", "#2b7bb9", "#2a7ab9", "#2979b8", "#2878b8", "#2777b7", "#2676b6", "#2574b6", "#2473b5", "#2372b4", "#2371b4", "#2270b3", "#216fb3", "#206eb2", "#1f6db1", "#1e6cb0", "#1d6bb0", "#1c6aaf", "#1c69ae", "#1b68ae", "#1a67ad", "#1966ac", "#1865ab", "#1864aa", "#1763aa", "#1662a9", "#1561a8", "#1560a7", "#145fa6", "#135ea5", "#135da4", "#125ca4", "#115ba3", "#115aa2", "#1059a1", "#1058a0", "#0f579f", "#0e569e", "#0e559d", "#0e549c", "#0d539a", "#0d5299", "#0c5198", "#0c5097", "#0b4f96", "#0b4e95", "#0b4d93", "#0b4c92", "#0a4b91", "#0a4a90", "#0a498e", "#0a488d", "#09478c", "#09468a", "#094589", "#094487", "#094386", "#094285", "#094183", "#084082", "#083e80", "#083d7f", "#083c7d", "#083b7c", "#083a7a", "#083979", "#083877", "#083776", "#083674", "#083573", "#083471", "#083370", "#08326e", "#08316d", "#08306b" ]35:T40a,temp_thresholds = [ { min: null, max: 0, label: "The surface is frozen" }, { min: 0, max: 8, label: "The temperature is dangerously freezing" }, { min: 8, max: 10, label: "The temperature is freezing like an ice bath and still dangerously cold" }, { min: 10, max: 12, label: "The temperature is a notch beyond freezing, but you might consider jumping in" }, { min: 12, max: 14, label: "The temperature is icy, but therapeutic and borderline swimmable" }, { min: 14, max: 18, label: "The temperature is chilly but swimmable" }, { min: 18, max: 20, label: "Pretty chilly in the normal summer way" }, { min: 20, max: 22, label: "Cool; near summer peak" }, { min: 22, max: 24, label: "The temperature is at the summer peak, (sort of) approaching the temperature of a pool" }, { min: 24, max: 100, label: "The temperature is near or beyond historical highs" } ]36:[{"id":"a748b72689f3ac823ccb418d73f73133e4ab54548d37bc646823ef463ed2056a8bb8883a3ded4e8e97ac8535663687ea8fd897c4f10b09b0da953d0979d2e1da","href":"https://static.observableusercontent.com/files/a748b72689f3ac823ccb418d73f73133e4ab54548d37bc646823ef463ed2056a8bb8883a3ded4e8e97ac8535663687ea8fd897c4f10b09b0da953d0979d2e1da","name":"signal-2022-08-19-220337 (1).jpeg","type":"image/jpeg","lastModified":1687115190922.093,"size":291831},{"id":"40e68fec0d6ee3580bdd4722f900c55ddc5f77791277537d3a7a6790c5c249bc7521902e8cc19f82fdc281f59b017fe381735b968bff6980139eb8e79fdb57d3","href":"https://static.observableusercontent.com/files/40e68fec0d6ee3580bdd4722f900c55ddc5f77791277537d3a7a6790c5c249bc7521902e8cc19f82fdc281f59b017fe381735b968bff6980139eb8e79fdb57d3","name":"signal-2022-08-20-123454.jpeg","type":"image/jpeg","lastModified":1687115190894.22,"size":354581},{"id":"b2b463b658a95e371580ffdb80689332ab0b63deaaeb33582c11b08dbf24e28b761f78d57e1fcbb892c89ba768fc5e83e0fe3ce17b9a49999225ee12b8de18c8","href":"https://static.observableusercontent.com/files/b2b463b658a95e371580ffdb80689332ab0b63deaaeb33582c11b08dbf24e28b761f78d57e1fcbb892c89ba768fc5e83e0fe3ce17b9a49999225ee12b8de18c8","name":"signal-2022-08-19-220350.jpeg","type":"image/jpeg","lastModified":1687115190975.828,"size":320707}] 2e:{"title":"What's it like on Lake Michigan on the North Side of Chicago right now?","cells":[{"id":0,"value":"$30","pinned":false,"hidden":false,"mode":"md"},{"id":1036,"value":"$31","pinned":false,"hidden":false,"mode":"ojs"},{"id":336,"value":"$32","pinned":false,"hidden":false,"mode":"html"},{"id":461,"value":"$33","pinned":false,"hidden":false,"mode":"md"},{"id":499,"value":"# Work-in-progress visual exploration\n\n_Only hardcore nerds may care to read beyond this point!_\n\nCharts that let us look at rolling windows for different measurements; still todo is an exploration of wave height vs. period.","pinned":false,"hidden":false,"mode":"md"},{"id":183,"value":"viewof windowSize = Inputs.select([0, 1, 3, 6, 24, 3 * 24], {\n label: \"Window size (hours)\",\n value: 0\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":800,"value":"schemeBlues = [\n lotsOfBlues[15],\n lotsOfBlues[40],\n lotsOfBlues[65],\n lotsOfBlues[90],\n lotsOfBlues[115],\n lotsOfBlues[125],\n lotsOfBlues[135]\n]","pinned":false,"hidden":false,"mode":"ojs"},{"id":831,"value":"$34","pinned":false,"hidden":false,"mode":"ojs"},{"id":757,"value":"chartZones = [\n { label: \"Dangerous\", temp: 12, description: \"Less than 12c\" },\n { label: \"Bracing\", temp: 14, description: \"12-14c\" },\n { label: \"“Therapeutic”\", temp: 16, description: \"14c-16c\" },\n { label: \"Tolerable\", temp: 18, description: \"16c-18c\" },\n { label: \"Swimmable\", temp: 20, description: \"20c-22c\" },\n { label: \"🤩\", temp: 22, description: \"20c-24c\" },\n { label: \":/\", temp: 24, description: \"24c+\" }\n].reverse()","pinned":false,"hidden":false,"mode":"ojs"},{"id":628,"value":"chartRangeStart = moment(date_range.get(\"max_date\", 0))\n .subtract(10, \"days\")\n .toDate()","pinned":false,"hidden":false,"mode":"ojs"},{"id":607,"value":"temperatureChartData = processed\n .params({ chartRangeStart })\n .filter(\n (d) =>\n aq.op.equal(d.parameter_label, \"Surface temp (degrees Celsius)\") &&\n d.date > chartRangeStart\n )","pinned":false,"hidden":false,"mode":"ojs"},{"id":501,"value":"# Rendering code","pinned":false,"hidden":false,"mode":"md"},{"id":272,"value":"function getThresholdLabel(value, thresholds) {\n return thresholds\n .sort((a, b) => b.min - a.min) // Find logic requires reverse sort\n .find(\n (threshold) =>\n (value >= threshold.min || !threshold.min) &&\n (value < threshold.max || !threshold.max)\n ).label;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":300,"value":"wave_height_thresholds = [\n {\n min: 2,\n max: null,\n label:\n \"The waves might be tall enough to surf on, the beach is definitely closed\"\n },\n {\n min: 1,\n max: 2,\n label: \"The waves are big and the beach might be closed\"\n },\n {\n min: -1,\n max: 0,\n label: \"The waves are fairly normal sized\"\n },\n {\n min: -3,\n max: -1,\n label: \"The waves are barely a ripple\"\n },\n {\n min: null,\n max: -3,\n label: \"The lake is a mirror\"\n }\n]","pinned":false,"hidden":false,"mode":"ojs"},{"id":268,"value":"wave_period_thresholds = [\n {\n min: 3,\n max: null,\n label: \"The waves are eerily calm\"\n },\n {\n min: 1,\n max: 3,\n label: \"The waves are stately and slow\"\n },\n {\n min: -1,\n max: 1,\n label: \"The waves have a rather normal pace\"\n },\n {\n min: -3,\n max: -1,\n label: \"The waves are rolling faster than normal\"\n },\n {\n min: null,\n max: -3,\n label: \"The waves are coming really fast\"\n }\n]","pinned":false,"hidden":false,"mode":"ojs"},{"id":295,"value":"$35","pinned":false,"hidden":false,"mode":"ojs"},{"id":503,"value":"# Data processing and utils","pinned":false,"hidden":false,"mode":"md"},{"id":310,"value":"summary = processed.groupby(\"parameter_label\").rollup({\n max: aq.op.max(\"value\"),\n min: aq.op.min(\"value\"),\n mean: aq.op.mean(\"value\"),\n lo: aq.op.quantile(\"value\", 0.1),\n median: aq.op.quantile(\"value\", 0.5),\n hi: aq.op.quantile(\"value\", 0.9),\n max_z: aq.op.max(\"z\"),\n min_z: aq.op.min(\"z\")\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":412,"value":"recentData = recent.objects()","pinned":false,"hidden":false,"mode":"ojs"},{"id":225,"value":"recent = processed\n .groupby(\"parameter_label\")\n .derive({\n last_value: (d) => aq.op.last_value(d.value),\n last_z: (d) => aq.op.last_value(d.z)\n })\n .groupby(\"parameter_label\", \"last_value\", \"last_z\")\n .rollup()","pinned":false,"hidden":false,"mode":"ojs"},{"id":149,"value":"Inputs.table(summary, { layout: true })","pinned":false,"hidden":false,"mode":"ojs"},{"id":442,"value":"date_range = dt.rollup({\n min_date: aq.op.min(\"date\"),\n max_date: aq.op.max(\"date\")\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":123,"value":"processed = dt\n .params({ windowSize, kelvin })\n .derive({\n value: (d) =>\n d.parameter_label === \"Surface temp (degrees Celsius)\"\n ? d.value - kelvin\n : d.value\n })\n .groupby(\"parameter_label\")\n .derive({\n rolling_average: aq.rolling(\n (d) => aq.op.average(d.value),\n [-windowSize, 0]\n ),\n z: (d) => (d.value - aq.op.mean(d.value)) / aq.op.stdev(d.value) || 0,\n dev: (d) => d.value - aq.op.median(d.value)\n })\n// .derive({ dev: d => d.x - op.median(d.x) })\n// .derive({ dev: d => d.dev / op.median(op.abs(d.dev)) || 0 })","pinned":false,"hidden":false,"mode":"ojs"},{"id":329,"value":"kelvin = 273.15","pinned":false,"hidden":false,"mode":"ojs"},{"id":112,"value":"dt = aq.from(data).orderby(\"date\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":102,"value":"data = responseData[0].parameters\n .map((param) =>\n param.observations.map((observation) => ({\n ...observation,\n moment: moment(observation.timestamp),\n date: moment(observation.timestamp).toDate(),\n parameter_id: param.parameter_id,\n parameter_label: parameterIds[param.parameter_id]\n }))\n )\n .flat()","pinned":false,"hidden":false,"mode":"ojs"},{"id":45,"value":"responseData = fetch(\n `https://seagull-api.glos.org/api/v1/obs?${querystring}`\n).then((response) => response.json())","pinned":false,"hidden":false,"mode":"ojs"},{"id":61,"value":"querystring = new URLSearchParams({\n startDate: start.format(\"YYYY-MM-DD\"),\n endDate: today.format(\"YYYY-MM-DD\"),\n obsDatasetId: 98,\n parameterId: Object.keys(parameterIds).join(\",\")\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":79,"value":"start = today.clone().subtract(\"months\", 3)","pinned":false,"hidden":false,"mode":"ojs"},{"id":63,"value":"today = moment()","pinned":false,"hidden":false,"mode":"ojs"},{"id":48,"value":"parameterIds = ({\n \"2949\": \"Surface temp (degrees Celsius)\",\n \"1644\": \"Mean wave period (seconds)\",\n \"1645\": \"Significant wave height (meters)\"\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":74,"value":"moment = require(\"moment\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":391,"value":"","pinned":false,"hidden":false,"mode":"html"}],"files":"$W36","stdlib":"1","titling":"auto"}