30:T5f4,# Exploring data with the Kepler Space Telescope In 2015, I read about NASA’s [Kepler Space Telescope](https://exoplanets.nasa.gov/keplerscience/). From 2009 to 2018, it stared out into deep space. It was one of the first instruments used to find **exoplanets**: planets circling stars other than our own. It turns out Kepler was parked at a [Lagrange point](https://www.youtube.com/watch?v=2LjGgJ4ST60), where the gravity of the Sun and the Earth balance each other. From this stable vantage point, it was measuring the brightness of stars, watching for the ones that got a little dimmer every so often. If a star’s brightness dips slightly on a regular schedule, that could be a sign there was a planet orbiting that star. Each time the planet between its star and us, it would block a bit of the star’s light. Astronomers studied the data for these periodic dimming events, calling each an **object of interest**. By correlating these measurements with observations from terrestrial telescopes, they could confirm whether they had truly found an exoplanet. The article I read said how the Kepler mission had found 8 Earth-like planets orbiting other stars. Being a nerd who grew up on science fiction (I mean, *please,* this is a *telescope* in *space!*), I downloaded the data and dug into it myself. I'd like to share my (very) amateur astronomy journey with you. _Cover thumbnail image credit: [NASA/Ames Research Center/Daniel Rutter](https://www.nasa.gov/image-feature/imagining-another-earth)_31:T524,The planets’ temperatures dropping as they get farther from their stars helps confirm we’re interpreting these values correctly. It’s also interesting to note that these planets all orbit fairly close to their stars. In [our solar system](https://solarsystem.nasa.gov/solar-system/our-solar-system/overview/), Mars is 1.5 AU from the Sun, and Uranus orbits out at 19.8 AU. Is that due to a bias in how Kepler detects planets? A sign that our star’s children are more scattered than is typical? Questions for another day... A little further digging tells me that Earth’s equilibrium temperature is 255 Kelvin (−1°F/−18°C). At first, that seems too cold, since water freezes well above that temperature. However, this only measures the theoretical temperature of a planet based on how much stellar radiation it absorbs. On Earth, the greenhouse effect raises our mean surface temperature to a much more comfortable 288 K (59°F/15°C). So what would Astronomer Goldilocks say is too hot? Or too big? 275 K is about the highest equilibrium temperature for water not to evaporate away, and above 1.75 Earth radii you’re starting to feel too much gravity or trying to stand a gas giant. When we remove planets above these thresholds, we are left with this:32:T66a,// Goldilocks scatter in a variable for custom placement and background color of the legend function goldilocks_scatter() { return Plot.plot({ marks: [ Plot.dot( objects_of_interest.filter((koi) => koi["Disposition"] == "Confirmed"), { x: "Orbit Size compared to Earth's", y: "Planetary Radius (Earths)", r: "Planetary Radius (Earths)", fill: "Equilibrium Temperature" } ), Plot.ruleY(objects_of_interest, { y: 1.0, strokeWidth: 0.5, stroke: "#444444" }), Plot.ruleX(objects_of_interest, { x: 1.0, strokeWidth: 0.5, stroke: "#444444" }), Plot.dot(objects_of_interest, { x: 1.0, y: 1.0, r: 1, fill: "#96c4db" }), Plot.text(objects_of_interest, { x: 1.0, y: 1.0, dx: 21, dy: -10, text: ["Earth"], fill: "#cccccc" }) ], color: { type: "linear", scheme: "buylrd", domain: [0, 1000], label: "Equilibrium Temperature (K)" }, x: { label: "Orbit Size (AU)", labelAnchor: "center", labelOffset: 45, tickSize: 3, ticks: 5, line: true }, y: { type: "log", label: "Radius (Earth radii)", labelAnchor: "center", labelOffset: 65, tickSize: 3, line: true }, marginLeft: 90, marginBottom: 65, marginRight: 35, inset: 20, style: { fontSize: "13px", color: "#e0e0e0", fontFamily: "'Open Sans', Lato, sans-serif", fontWeight: 550, background: "#222222" } }); }33:Ta11,// Goldilocks scatter in a variable for custom placement and background color of the legend function goldilocks_scatter_filtered() { const filteredSet = objects_of_interest .filter((koi) => koi["Disposition"] == "Confirmed") .filter((koi) => koi["Planetary Radius (Earths)"] <= 1.75) .filter((koi) => koi["Equilibrium Temperature"] <= 275.0); return Plot.plot({ marks: [ Plot.dot(filteredSet, { x: "Orbit Size compared to Earth's", y: "Planetary Radius (Earths)", r: 5, fill: golditoggle ? (d) => d["Kepler Object Name"] != "Kepler-395 c" ? "Goldilocks" : "Other" : "Equilibrium Temperature" }), // Label to right of mark (would be better to do it using the axis range, but I know all the marks) Plot.text( filteredSet .filter((koi) => golditoggle) .filter((koi) => [ "442 b", "440 b", "296 f", "186 f", "296 e", "438 b", "-62 e" ].some((x) => x === koi["Kepler Object Name"].slice(-5)) ), { x: "Orbit Size compared to Earth's", y: "Planetary Radius (Earths)", text: "Kepler Object Name", dx: 9, textAnchor: "start" } ), // Label to left of mark Plot.text( filteredSet .filter((koi) => golditoggle) .filter((koi) => "Kepler-62 f" === koi["Kepler Object Name"]), { x: "Orbit Size compared to Earth's", y: "Planetary Radius (Earths)", text: "Kepler Object Name", dx: -10, textAnchor: "end" } ) ], color: golditoggle ? { domain: ["Other", "Goldilocks"], range: ["#9498a0", "#ff725c"] } : { type: "linear", scheme: "buylrd", domain: [0, 1000], label: "Equilibrium Temperature (K)" }, x: { label: "Orbit Size (AU)", labelAnchor: "center", labelOffset: 45, tickSize: 3, ticks: 5, line: true, }, y: { type: "log", label: "Radius (Earth radii)", labelAnchor: "center", labelOffset: 65, tickSize: 3, line: true }, marginLeft: 90, marginBottom: 65, marginRight: 35, inset: 20, style: { fontSize: "13px", color: "#e0e0e0", fontFamily: "'Open Sans', Lato, sans-serif", fontWeight: 550, background: "#222222" } }); }34:T469,35:[{"id":"ca097a105e3182f2705c6f0d9a8b9a7cb03a3cac8ea3b88dc3469bf078cf0fed0f982e36630a37fffb772af215cd756d25e21e8daa2703cf84f6871956a3f40b","href":"https://static.observableusercontent.com/files/ca097a105e3182f2705c6f0d9a8b9a7cb03a3cac8ea3b88dc3469bf078cf0fed0f982e36630a37fffb772af215cd756d25e21e8daa2703cf84f6871956a3f40b","name":"KOI March 2015.csv","type":"text/csv","lastModified":1674347597040.603,"size":1339933},{"id":"a2092908f84e2e53b1c45ee792c72ab686a24fe2021057e1337f9b919c5b3fad8ae06fdae1cb25e05265a2cfdd6ee01b84408ffd8a69a5cec4629282f2708eff","href":"https://static.observableusercontent.com/files/a2092908f84e2e53b1c45ee792c72ab686a24fe2021057e1337f9b919c5b3fad8ae06fdae1cb25e05265a2cfdd6ee01b84408ffd8a69a5cec4629282f2708eff","name":"New Project (11).jpg","type":"image/jpeg","lastModified":1674511460646.103,"size":68185},{"id":"48a0b592ea91f77f73b63c05249777201d8edc5f31a0d00f083f434105a3255f19a287dfb1b5a1d55f17918d3fc866518752391a711e421f74595a4d2513948e","href":"https://static.observableusercontent.com/files/48a0b592ea91f77f73b63c05249777201d8edc5f31a0d00f083f434105a3255f19a287dfb1b5a1d55f17918d3fc866518752391a711e421f74595a4d2513948e","name":"Kepler telescope photometer@1.jpg","type":"image/jpeg","lastModified":1674517377912.555,"size":497916}] 2e:{"title":"Exploring data with the Kepler Space Telescope","cells":[{"id":0,"value":"$30","pinned":false,"hidden":false,"mode":"md"},{"id":464,"value":"

","pinned":false,"hidden":false,"mode":"html"},{"id":486,"value":"## By the numbers","pinned":false,"hidden":false,"mode":"md"},{"id":433,"value":"NASA had published several tables of data, including the full set of observations of brightness of 170,000 stars for several years. I was not trying to replicate the science, but rather to understand what had been done, so I downloaded one of the most processed data sets: the table listing the objects of interest, or as I thought of them, “maybe-planets.”\n\nThis data set had nearly 8,000 of objects of interest, each labeled as one of three possibilities:\n\n- Candidate — maybe a planet, but not verified by other telescopes yet\n- False positives, where the periodic dimming could not be explained as an exoplanet\n- Confirmed exoplanets","pinned":false,"hidden":false,"mode":"md"},{"id":797,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":10,"value":"Plot.plot({\n marks: [\n Plot.barX( \n objects_of_interest, \n Plot.groupY( { x: \"count\" }, { y: \"Disposition\", fill: \"Disposition\", sort: {y: \"x\", reverse: true} } ),\n ),\n Plot.text( \n objects_of_interest, \n Plot.groupY( { x: \"count\", text: \"count\" }, { y: \"Disposition\", dx: 8, textAnchor: \"start\" } ),\n ),\n ],\n x: {\n label: \"Number of Objects of Interest\",\n labelAnchor: \"center\",\n labelOffset: 35,\n tickPadding: 5,\n markAnchor: \"left\",\n ticks: 0,\n },\n y: {\n label: null,\n tickSize: 0,\n },\n color: {\n domain: [\"False Positive\", \"Candidate\", \"Confirmed\"],\n range: [\"#9498a0\", \"#4269d0\", \"#efb118\"]\n },\n marginLeft: 120,\n marginBottom: 65,\n marginRight: 75,\n marginTop: 35,\n style: {\n fontSize: \"13px\",\n color: \"#e0e0e0\",\n fontFamily: \"'Open Sans', Lato, sans-serif\",\n fontWeight: 550,\n background: \"#222222\"\n },\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":795,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":446,"value":"So while Kepler had found nearly 1,000 confirmed exoplanets, NASA had announced they had found only 8 Earth-like planets. Fewer than 1% of its discoveries were of planets that could find life hospitable to creatures something like us. I would come back to dig into that a bit later, but first...","pinned":false,"hidden":false,"mode":"md"},{"id":467,"value":"

","pinned":false,"hidden":false,"mode":"html"},{"id":474,"value":"## Where were we looking?","pinned":false,"hidden":false,"mode":"md"},{"id":435,"value":"With 1,000 exoplanets found so far and only 8 even vaguely like ours, I wondered how much of the sky we were studying. Were these exoplanets scattered across the whole sky? A hemisphere? A corner?\n\nI did some research (ok, I googled...) and discovered that just like we have latitude and longitude for the surface of the Earth, **declination** and **right ascension** are a positioning system for looking out into space. Like latitude, declination runs from −90 to +90 degrees, and right ascension wraps around “horizontally” like longitude, from 0 to 360 degrees.\n\nI plotted the objects of interest in a scatter plot, much like a Mercator projection, with right ascension was on the x-axis and declination on the y. I discovered that Kepler was only staring at this tiny postage stamp part of the sky.","pinned":false,"hidden":false,"mode":"md"},{"id":799,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":217,"value":"Plot.plot({\n marks: [\n Plot.dot(objects_of_interest, {x: \"Right Ascension\", y: \"Declination\", r: 1, fill: \"Disposition\" }),\n Plot.ruleY(objects_of_interest, { y: 0, strokeWidth: 1, stroke: \"#666666\" }),\n ],\n color: {\n domain: [\"False Positive\", \"Candidate\", \"Confirmed\"],\n range: [\"#9498a0\", \"#4269d0\", \"#efb118\"]\n },\n x: {\n domain: [0, 360],\n label: \"Right Ascension\",\n labelAnchor: \"center\",\n labelOffset: 45,\n nice: false,\n tickFormat: d => `${d}º`,\n tickSize: 3,\n },\n y: {\n domain: [-90, 90],\n label: \"Declination\",\n labelAnchor: \"center\",\n labelOffset: 50,\n tickFormat: d => `${d}º`,\n tickSize: 3,\n },\n width: 800,\n height: 400,\n marginLeft: 75,\n marginBottom: 65,\n marginRight: 35,\n style: {\n fontSize: \"13px\",\n color: \"#e0e0e0\",\n fontFamily: \"'Open Sans', Lato, sans-serif\",\n fontWeight: 550,\n background: \"#222222\"\n },\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":801,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":438,"value":"But that cross-shaped distribution caught my attention. What was it about Kepler that captured that pattern in the sky?","pinned":false,"hidden":false,"mode":"md"},{"id":478,"value":"

","pinned":false,"hidden":false,"mode":"html"},{"id":477,"value":"## That's no Nikon there\n\nI zoomed in, keeping only the part of the sky where the telescope was focused, and the shape of what I saw truly surprised me.","pinned":false,"hidden":false,"mode":"md"},{"id":804,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":619,"value":"

${ccd_scatter()}

${ccd_scatter().legend(\"color\")}

","pinned":false,"hidden":false,"mode":"html"},{"id":807,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":454,"value":"Why was Kepler capturing this window-frame grid of squares? I researched further (well, I googled more...) and found this photo.","pinned":false,"hidden":false,"mode":"md"},{"id":810,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":360,"value":"

$\"Photo$

Image credit: NASA/Ball Aerospace

","pinned":false,"hidden":false,"mode":"html"},{"id":813,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":462,"value":"This is a picture of the imaging array inside the Kepler Space Telescope, taken before it was installed. Each of the blue rectangles is a **charged coupling device** (CCD). It’s the same kind of technology we all carry with us in our cell phones to capture digital photos.\n\nIf you look closely at the four black squares in the corner of the array, you can see tiny blue squares inside them. Those are low-resolution CCDs. The smaller images captured with these can be transmitted more quickly to Earth than the full images. Astronomers use these low-res pictures to ensure that Kepler was still pointed accurately before downloading the larger, more expensive ones.","pinned":false,"hidden":false,"mode":"md"},{"id":869,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":557,"value":"### Even more trivial...","pinned":false,"hidden":false,"mode":"md"},{"id":457,"value":"

\n $\"Photo$ \n

Ball Aerospace was one of the contractors that worked on the Kepler Space Telescope. You can see the same Ball logo on the lab coat in the photo with the CCDs and on this home canning jar because Ball Corporation makes both “metal packaging for beverages, foods and household products” and “aerospace and other technologies.”

","pinned":false,"hidden":false,"mode":"html"},{"id":551,"value":"

","pinned":false,"hidden":false,"mode":"html"},{"id":563,"value":"## Where should we go?","pinned":false,"hidden":false,"mode":"md"},{"id":565,"value":"Now that I had some idea how Kepler worked, I was wondering if I could find those 8 Earth-like planets from among the 993 confirmed objects of interest. Some astronomers call them **Goldilocks planets**, because they are the planets that are not too big and not too small, not too hot and not too cold. I spread these confirmed planets out in a scatter plot of orbit size vs. planet size, and also colored by the temperature.","pinned":false,"hidden":false,"mode":"md"},{"id":816,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":712,"value":"

${goldilocks_scatter()}

\n \n

${goldilocks_scatter().legend(\"color\")}

","pinned":false,"hidden":false,"mode":"html"},{"id":819,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":744,"value":"$31","pinned":false,"hidden":false,"mode":"md"},{"id":822,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":1181,"value":"goldilocks_scatter_filtered()","pinned":false,"hidden":false,"mode":"ojs"},{"id":848,"value":"viewof golditoggle = {\n const golditoggle = document.createElement(\"div\");\n \n const e = document.createElement(\"span\");\n e.setAttribute(\"style\", \"font-family: 'Open Sans', Lato, sans-serif; font-size: 14px;\");\n const t = document.createTextNode(\"Highlight first 8 Goldilocks planets\");\n e.appendChild(t);\n golditoggle.appendChild(e);\n \n const b = document.createElement(\"input\");\n b.setAttribute(\"id\", \"hilite\");\n b.setAttribute(\"type\", \"checkbox\");\n b.setAttribute(\"class\", \"cm-toggle\");\n golditoggle.appendChild(b);\n\n golditoggle.value = false;\n b.onclick = (e => { \n b.dispatchEvent(new Event(\"input\"));\n golditoggle.value = b.checked;\n } );\n\n return golditoggle;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":825,"value":" ","pinned":false,"hidden":false,"mode":"html"},{"id":830,"value":"Nine planets. As of the release of this data set from Kepler, NASA had identified 8 Earth-like planets. How many of them did we find with some back-of-the-napkin math? To see the overlap between those 8 planets and these 9, toggle the switch.","pinned":false,"hidden":false,"mode":"md"},{"id":964,"value":"

","pinned":false,"hidden":false,"mode":"html"},{"id":963,"value":"## Conclusion\n\nAs downloaded, the Kepler Objects of Interest data set is an inscrutable table of fields with obscure meanings. Using visual exploration of the data, I was able to move through and understand about 8,000 maybe-planets, finding one thing out, doing research, then asking more questions. I learned about:\n- The early success of the Kepler mission\n- How astronomers divide their observations\n- Using the brightness of stars to find exoplanets\n- Construction of the telescope\n- Using the “Goldilocks” parameters to isolate Earth-like exoplanets — and getting lucky enough to find the first 8 Earth-like planets (and one mistaken one)\n\nBetter still, I have been able to use this insight to teach thousands of other people about the power of visual analytics, using this story as a living example of effective storytelling with charts and graphs. I hope you've gotten something out of this exploration too.","pinned":false,"hidden":false,"mode":"md"},{"id":960,"value":"

","pinned":false,"hidden":false,"mode":"html"},{"id":175,"value":"## Appendix","pinned":false,"hidden":false,"mode":"md"},{"id":345,"value":"function ccd_scatter() {\n return Plot.plot({\n marks: [\n Plot.dot(objects_of_interest, {x: \"Right Ascension\", y: \"Declination\", r: 1.5, fill: \"Disposition\" }),\n ],\n color: {\n domain: [\"False Positive\", \"Candidate\", \"Confirmed\"],\n range: [\"#9498a0\", \"#4269d0\", \"#efb118\"]\n },\n x: {\n label: \"Right Ascension\",\n labelAnchor: \"center\",\n labelOffset: 45,\n nice: false,\n tickFormat: d => `${d}º`,\n tickSize: 3,\n },\n y: {\n label: \"Declination\",\n labelAnchor: \"center\",\n labelOffset: 50,\n tickFormat: d => `${d}º`,\n tickSize: 3,\n },\n marginLeft: 75,\n marginBottom: 65,\n marginRight: 35,\n inset: 20,\n width: 800,\n height: 450,\n style: {\n fontSize: \"13px\",\n color: \"#e0e0e0\",\n fontFamily: \"'Open Sans', Lato, sans-serif\",\n fontWeight: 550,\n background: \"#222222\"\n },\n })\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":687,"value":"$32","pinned":false,"hidden":false,"mode":"ojs"},{"id":782,"value":"$33","pinned":false,"hidden":false,"mode":"ojs"},{"id":846,"value":"$34","pinned":false,"hidden":false,"mode":"html"},{"id":876,"value":"objects_of_interest = await FileAttachment(\"KOI March 2015.csv\").csv({typed: \"auto\"})","pinned":false,"hidden":false,"mode":"ojs"}],"files":"$W35","stdlib":"1","titling":"auto"}