2b:null 2c:["$","title",null,{"children":"GitHub Issues Survival Analysis | Stas | Observable"}] 31:T625,In my previous notebook [@stas-sl/clickhouse-playground](https://observablehq.com/@stas-sl/clickhouse-playground) I've already demonstrated how to connect and use the [ClickHouse Playground](https://clickhouse.com/docs/en/getting-started/playground), which, in case if you don't know, is a collection of datasets stored in ClickHouse database and accessible via HTTP. Here, inspired by great work in [GitHub Burndown](https://observablehq.com/@tmcw/github-burndown) and [Everything You Always Wanted To Know About GitHub (But Were Afraid To Ask)](https://ghe.clickhouse.tech/), I'd like to, in a sense, combine ideas from both of them. Compared to the mentioned notebook, where data is retrieved via GitHub API, that could took a long time, in case if a repo has thousands of issues, in this notebook I will query the data from `github_events` table, that will take only a few seconds even for large repositories. `github_events` stores all the data from [www.gharchive.org](https://www.gharchive.org/) site, but instead of JSON, which could be quite slow, if you need to aggregate the data over large time periods, in ClickHouse it is stored in column-based form for fast and efficient retrieval. It includes _4.5 billon_ events for _20+_ event types starting from _2011_ year and is regularly ([every 10 minutes](https://github.com/ClickHouse/github-explorer/issues/6#issuecomment-996201290)) updated (at least at the time of writing). In this notebook I'll be using only small part (_145 million_ rows) of this dataset containing single `event_type = 'IssuesEvent'`.32:T6b5,{ const _data = await data[0]; const plot = await embed( { $schema: "https://vega.github.io/schema/vega-lite/v5.json", width: width - 30, height: 300, autosize: "fit-x", data: { values: _.flatten( _data.slice(0).map((x) => _.zip(x.date, x.count).map(([date, count]) => ({ created_at: Date.parse(x.created_at), date, count })) ) ) }, title: "Open issues colored by creation date", mark: "area", encoding: { x: { field: "date", type: "temporal", axis: { grid: false, title: "", tickCount: 10 } }, y: { aggregate: "sum", field: "count", axis: { title: "", tickCount: 5 } }, color: { field: "created_at", type: "quantitative", scale: { scheme: { name: "turbo", extent: [0.1, 0.9] }, reverse: true }, legend: { values: d3 .scaleUtc() .domain(d3.extent(_data.map((d) => Date.parse(d.created_at)))) .ticks(5) .map((x) => x.getTime()), formatType: "time", format: "%b\u2019%y", orient: "top", gradientLength: 300, title: "Created at" } }, order: { field: "created_at" } } }, _.merge(_.cloneDeepWith(vegaConfig), { config: { legend: { labelFontSize: 12 } } }) ); invalidation.then(() => plot.value.finalize()); return plot; }33:T521,{ const _data = await data[1]; const countMax = d3.quantile( _data.map((x) => x.count), 0.995 ) * 1.1; const domain = [-countMax, countMax]; const plot = await vl .markBar({ interpolate: "linear", width: Math.max(1, (width / _data.length) * 0.4), clip: true }) .data(_data) .transform( vl .calculate("datum.state === 'closed' ? -datum.count : datum.count") .as("count") ) .encode( vl.x().fieldT("date").axis({ grid: false, title: false, tickCount: 10 }), vl .y() .fieldQ("count") .title("") .scale({ domain }) .axis({ tickCount: 5, gridColor: { condition: { test: "datum.value === 0", value: "black" }, value: "#ddd" } }), vl .color() .fieldN("state") .scale({ range: ["#ee6666", "#91cc75"] }) .legend({ orient: "top", title: "", values: ["opened", "closed"] }) ) .width(width) .height(300) .autosize({ type: "fit-x", contains: "padding" }) .title("Opened vs closed issues") .config(vegaConfig.config) .render({ renderer: "svg" }); invalidation.then(() => plot.value.finalize()); return plot; }34:T4d5,{ const _data = data[2]; if (histOpts.state.length === 0) return html``; const ageFilter = d3.quantile( _data.map((x) => x.age_days), histOpts.ageFilter ); const plot = await vl .markBar() .data( _data.filter( (x) => x.age_days <= ageFilter && histOpts.state.includes(x.state) ) ) .encode( vl .x() .fieldQ("age_days") .bin({ anchor: 0, maxbins: histOpts.maxbins, minstep: 1 }) .axis({ grid: false, title: "Age (days)" }), vl.y().aggregate("count").title("Count").axis({ tickCount: 5 }), vl .color() .fieldN("state") .scale({ domain: histOpts.state, range: histOpts.state.map( (x) => ({ open: "#91cc75", closed: "#ee6666" }[x]) ) }) .legend({ orient: "top", title: "" }), vl.order().field("state") ) .width(width) .height(300) .autosize({ type: "fit-x", contains: "padding" }) .title("Issues age distribution") .config(vegaConfig.config) .render({ renderer: "svg" }); invalidation.then(() => plot.value.finalize()); return plot; }35:T753,{ let _data = data[2]; const KMtable = aq .from(_data) .params({ splitKM }) .derive({ age_days: (d) => +d.age_days, group: (d, $) => $.splitKM ? aq.substring(d.created_at, 0, 4) : "all" }) .groupby("group", "age_days") .rollup({ closed: (d) => aq.sum(d.is_closed), open: (d) => aq.sum(!d.is_closed) }) .groupby("group") .orderby(aq.desc("age_days")) .derive({ cum_open: aq.rolling((d) => aq.sum(d.open + d.closed)) }) .orderby("age_days") .derive({ hazard: (d) => d.closed / d.cum_open }) .derive({ survival: aq.rolling((d) => aq.product(1 - d.hazard)) }) .derive({ survival: (d) => aq.lag(d.survival) || 1 }) .select(aq.not("open")) .objects(); const plot = await embed( { $schema: "https://vega.github.io/schema/vega-lite/v5.json", width: width - 30, height: 300, autosize: "fit-x", data: { values: KMtable }, title: "Kaplan-Meier survival curve", mark: { type: "line" }, encoding: { x: { field: "age_days", type: "quantitative", axis: { grid: false, title: "Age (days)", tickCount: 10 } }, y: { field: "survival", type: "quantitative", axis: { title: "Survival probability", tickCount: 10 } }, color: { field: "group", type: "nominal", scale: splitKM ? { scheme: "turbo", reverse: true } : { scheme: "tableau10" }, legend: { orient: "top", title: "created at year", symbolStrokeWidth: 10 } } } }, vegaConfig ); invalidation.then(() => plot.value.finalize()); return htl.html`${plot}${Inputs.table(KMtable)}`; }36:T1077,data = { // for the first run (when you just open the notebook) use cached data from attachments // to reduce server load, and to make sure this notebooks will render at least once even // if server will go down some day if (mutable firstRun) { mutable firstRun = false; return await FileAttachment("data@4.json").json(); } const commonSQL = `with issue_events as ( select created_at, number, toString(action) as action from github_events where event_type = 'IssuesEvent' and repo_name in (${repoFormatted}) order by created_at ), issue_opened_events as ( select created_at, number, action from issue_events where action in ('opened', 'reopened') union all select toDateTime(0), number, '' from issue_events where number not in (select number from issue_events where action in ('opened', 'reopened')) ), issue_closed_events as ( select created_at, number, action from issue_events where action = 'closed' ), dates as ( select distinct toDate(created_at) as date from issue_events order by date with fill ), open_intervals as ( select o.number, o.created_at as opened_at, c.created_at as closed_at, dateDiff('day', o.created_at, c.created_at) as dur from issue_opened_events o asof left join issue_closed_events c on o.number = c.number and c.created_at > o.created_at order by number ), open_intervals_full as ( select number, min(opened_at) as opened_at, max(if(oi.closed_at = toDateTime(0), toDateTime('2100-01-01'), oi.closed_at)) as _closed_at, if(_closed_at = toDateTime('2100-01-01'), toDateTime(0), _closed_at) as closed_at, dateDiff('day', opened_at, closed_at) as age from open_intervals oi group by number )`; return Promise.all([ client.query( commonSQL + `, issues_vs_date as ( select date, oi.number as number, oi.opened_at as opened_at, oi.closed_at as closed_at, oif.opened_at as opened_at_final, oif.closed_at as closed_at_final, dateDiff('day', oi.opened_at, date) as age, dateDiff('day', oif.opened_at, oif.closed_at) as age_final, toDate(oif.closed_at) = date as is_closed_today from open_intervals oi join open_intervals_full oif on oi.number = oif.number cross join dates d where toDate(oi.opened_at) <= d.date and (toDate(oi.closed_at) >= d.date or toDate(oi.closed_at) = toDate(0)) ), counts as ( select date, toDateOrDefault(toStartOfInterval(if(opened_at_final = toDateTime(0), (select min(date) from dates), opened_at_final), interval 1 month)) as created_at, uniqExact(number) as count from issues_vs_date ivd group by date, created_at order by date, created_at ) select created_at, groupArray(date) as date, groupArray(count) as count from counts group by created_at; ` ), client.query( commonSQL + `, opened_count as ( select if(toDate(opened_at) = toDate(0), (select min(date) from dates), toDate(opened_at)) as opened_at, count(*) as count from open_intervals_full group by opened_at ), closed_count as ( select toDate(closed_at) as closed_at, count(*) as count from open_intervals_full group by closed_at ) select * from ( select toDate(opened_at) as date, if(date > (select min(date) from dates), count, 0) as count, sum(count) over (order by date) as cum_count, 'opened' as state from opened_count union all select toDate(closed_at) as date, count, sum(count) over (order by date) as cum_count, 'closed' as state from closed_count where date > toDate(0) ) order by date; ` ), client.query( commonSQL + `select number, opened_at as created_at, dateDiff('day', opened_at, if(closed_at = 0, now(), closed_at)) as age_days, closed_at != 0 as is_closed, if(is_closed, 'closed', 'open') as state from open_intervals_full where opened_at > 0 order by number; ` ) ]); }37:[{"id":"f80cd87e27828633921b29c5b8efd19e59d46f168ac9b3fa47e52607541601942770399e848124ca4a717849735ff7114d6dc768f0739fa603fdc623f1163603","href":"https://static.observableusercontent.com/files/f80cd87e27828633921b29c5b8efd19e59d46f168ac9b3fa47e52607541601942770399e848124ca4a717849735ff7114d6dc768f0739fa603fdc623f1163603","name":"data@4.json","type":"application/json","lastModified":1650149844813.792,"size":2615519},{"id":"2cf6356c1962371a71a8c179e7a0c500fb9eabce4bd5d0bc0b526ae8a8bdda256b6b0b927a891bf3ec91e2aa5cc5316b68822663f87b5cecc37947c516fa0c80","href":"https://static.observableusercontent.com/files/2cf6356c1962371a71a8c179e7a0c500fb9eabce4bd5d0bc0b526ae8a8bdda256b6b0b927a891bf3ec91e2aa5cc5316b68822663f87b5cecc37947c516fa0c80","name":"bokeh.png","type":"image/png","lastModified":1650139240143.542,"size":266904},{"id":"35f5f7b88603b006d2f421b4be11837ab69fe3dd249738051513488109525b3dfe836a5d9f23cbb8a920e62d2721ceb05bfc7009dd90ccc000641a002ab144b9","href":"https://static.observableusercontent.com/files/35f5f7b88603b006d2f421b4be11837ab69fe3dd249738051513488109525b3dfe836a5d9f23cbb8a920e62d2721ceb05bfc7009dd90ccc000641a002ab144b9","name":"pytorch.png","type":"image/png","lastModified":1650139239949.129,"size":146607},{"id":"62fa9d538ee464ea7fb42e0a9763e230410e31cec55bc556de4bb8d08592f04f78eaf61909547956e1f32f7ae955fd96eb7cfe0eb66a989d430e1524b422a77b","href":"https://static.observableusercontent.com/files/62fa9d538ee464ea7fb42e0a9763e230410e31cec55bc556de4bb8d08592f04f78eaf61909547956e1f32f7ae955fd96eb7cfe0eb66a989d430e1524b422a77b","name":"seaborn.png","type":"image/png","lastModified":1650139240084.772,"size":191931},{"id":"3316d804bd13161215a11eb05854c20b97416fa3809b2df0b53cb536381c96cb0667f2b0f26bbb348762127acdd3769060704c4f4c457cce49dcee3af256292e","href":"https://static.observableusercontent.com/files/3316d804bd13161215a11eb05854c20b97416fa3809b2df0b53cb536381c96cb0667f2b0f26bbb348762127acdd3769060704c4f4c457cce49dcee3af256292e","name":"altair.png","type":"image/png","lastModified":1650139239810.455,"size":119439},{"id":"2d6415687c681384735b732dd90bedbd45d5a578b00cd292a9c7b7a2f882f23ec6ce8f305ccdc6caef8457f7a80be1e47cc71bc0b75ee21f8d25d1e3c5c3f53e","href":"https://static.observableusercontent.com/files/2d6415687c681384735b732dd90bedbd45d5a578b00cd292a9c7b7a2f882f23ec6ce8f305ccdc6caef8457f7a80be1e47cc71bc0b75ee21f8d25d1e3c5c3f53e","name":"observable-feedback.png","type":"image/png","lastModified":1650139239780.661,"size":76098},{"id":"3a0818952f6c7c559dc81bfd7826a1d96787156a3e26b88e6763375befbbdb63afab9e41cbbdbba62e8dee479db369bbf8c9e70e4e5ab95f37ea82829246a454","href":"https://static.observableusercontent.com/files/3a0818952f6c7c559dc81bfd7826a1d96787156a3e26b88e6763375befbbdb63afab9e41cbbdbba62e8dee479db369bbf8c9e70e4e5ab95f37ea82829246a454","name":"observable-plot.png","type":"image/png","lastModified":1650139239808.317,"size":103257},{"id":"299058200ddf8190286d9c8bd262f22c4a2202639d970416112e115bb037e3c80a6c14a3247a5e128ce593b257407253afbe532799e1884de118c16b3fc4be08","href":"https://static.observableusercontent.com/files/299058200ddf8190286d9c8bd262f22c4a2202639d970416112e115bb037e3c80a6c14a3247a5e128ce593b257407253afbe532799e1884de118c16b3fc4be08","name":"matplotlib.png","type":"image/png","lastModified":1650139240216.501,"size":288986},{"id":"a984a9f41a823453b3a98328ef0ced980371558e20bce9cc2a677070631110bb9e5c5b351848af9eafb9bbd4cba4e942e0dc3848eb475e1c6ee0fdfa48450a72","href":"https://static.observableusercontent.com/files/a984a9f41a823453b3a98328ef0ced980371558e20bce9cc2a677070631110bb9e5c5b351848af9eafb9bbd4cba4e942e0dc3848eb475e1c6ee0fdfa48450a72","name":"tensorflow.png","type":"image/png","lastModified":1650139240213.84,"size":257736},{"id":"7eade01c629cc2818fc929cfa9d4f753be94cddc4fb3e1afac7b4ec8666ea00c98032d7fcaf11052b2ff8a95f4f008686a41154a802f300c6996413e6daca5e4","href":"https://static.observableusercontent.com/files/7eade01c629cc2818fc929cfa9d4f753be94cddc4fb3e1afac7b4ec8666ea00c98032d7fcaf11052b2ff8a95f4f008686a41154a802f300c6996413e6daca5e4","name":"echarts.png","type":"image/png","lastModified":1650139240117.665,"size":220112},{"id":"1efd905400fcfa563592605130c4066c80c0e611906e7e0ed30f1bd785efd450f59d31a2514846ca4236d149f2fe86688a2a3332763385f59790f1da864863bc","href":"https://static.observableusercontent.com/files/1efd905400fcfa563592605130c4066c80c0e611906e7e0ed30f1bd785efd450f59d31a2514846ca4236d149f2fe86688a2a3332763385f59790f1da864863bc","name":"vega.png","type":"image/png","lastModified":1650139239867.38,"size":105733},{"id":"a757fead673c270ea8378e71086a17c1507334aee70c93f4c5764290d42110287f917ddc7edae16ff00ef0a6938b86979283dde29e552f2d9b4d2e6b6d7303e5","href":"https://static.observableusercontent.com/files/a757fead673c270ea8378e71086a17c1507334aee70c93f4c5764290d42110287f917ddc7edae16ff00ef0a6938b86979283dde29e552f2d9b4d2e6b6d7303e5","name":"pandas.png","type":"image/png","lastModified":1650139240143.954,"size":266442},{"id":"f70f3648bd33637b7154143efdf11743cde54d2d1281347a0adffa1484be92e6c235253b91a64cc2d6a4b1084eace60ec5b9946be8a02a3389fd51ac219a25c0","href":"https://static.observableusercontent.com/files/f70f3648bd33637b7154143efdf11743cde54d2d1281347a0adffa1484be92e6c235253b91a64cc2d6a4b1084eace60ec5b9946be8a02a3389fd51ac219a25c0","name":"react.png","type":"image/png","lastModified":1650139240088.28,"size":240696},{"id":"eebe12bf5d11e901416e8d85363a810e7fbd0196dc2088dd31c1e1a6f77e834d6800db84ad60e2eb871f5ed6fdd6d38cf81b28ac756e140acaf9c32d1aa9c4e5","href":"https://static.observableusercontent.com/files/eebe12bf5d11e901416e8d85363a810e7fbd0196dc2088dd31c1e1a6f77e834d6800db84ad60e2eb871f5ed6fdd6d38cf81b28ac756e140acaf9c32d1aa9c4e5","name":"clickhouse.png","type":"image/png","lastModified":1650139240144.657,"size":212702},{"id":"9bff51cb1e6b3bc39c45dc703017c4c731b9300d4353f109a94ec5c06fd68ed58e911b49dc5f9d6a42c4f5f7dfe719ff1cb3448eeffee458225de439ae27d987","href":"https://static.observableusercontent.com/files/9bff51cb1e6b3bc39c45dc703017c4c731b9300d4353f109a94ec5c06fd68ed58e911b49dc5f9d6a42c4f5f7dfe719ff1cb3448eeffee458225de439ae27d987","name":"plotly.js.png","type":"image/png","lastModified":1650139240008.761,"size":137024},{"id":"8cefa4f1fe0a5e11da72213b9696e339b2d952cfa2221ee8a49f22a60d8f16b9f83676f9ba8378632a9ce19fb0ac192ea35498f547b73d5ece964d2ad0435c3b","href":"https://static.observableusercontent.com/files/8cefa4f1fe0a5e11da72213b9696e339b2d952cfa2221ee8a49f22a60d8f16b9f83676f9ba8378632a9ce19fb0ac192ea35498f547b73d5ece964d2ad0435c3b","name":"duckdb.png","type":"image/png","lastModified":1650139239943.231,"size":119554},{"id":"05ffbe86b5f1228d96ab50938609462fce575969e28b52e7f198ff8c2fe94fb47f8f4bb25d76333d10939fb1855e76f8003da0fb4fac9f1817b1475026d9a839","href":"https://static.observableusercontent.com/files/05ffbe86b5f1228d96ab50938609462fce575969e28b52e7f198ff8c2fe94fb47f8f4bb25d76333d10939fb1855e76f8003da0fb4fac9f1817b1475026d9a839","name":"vega-lite.png","type":"image/png","lastModified":1650139240005.529,"size":159666},{"id":"1600ad5609d02771906d2b6d8e8f03af623523523a2ef5f06d7820e1c8ac000529d9dea47f3122a857a17a9afce1082deb08d3bd5e752dc3d42beb4367d0c580","href":"https://static.observableusercontent.com/files/1600ad5609d02771906d2b6d8e8f03af623523523a2ef5f06d7820e1c8ac000529d9dea47f3122a857a17a9afce1082deb08d3bd5e752dc3d42beb4367d0c580","name":"vue.png","type":"image/png","lastModified":1650139239907.182,"size":131689}] 2f:{"id":"ba0a06c04cc3f0b6","slug":"github-issues-survival-analysis","title":"GitHub Issues Survival Analysis","thumbnail":"d18b5777bb94b0bbf2d1bfafd838615fdb81b546d7cdfd4b5340b2a61436f2d2","access_level":"public","create_time":"$D2022-04-14T20:38:35.315Z","update_time":"$D2022-09-21T19:36:26.258Z","view_time":null,"first_public_time":"$D2022-04-16T20:53:39.755Z","fork":true,"stars":5,"starred":false,"owner":{"type":"team","id":"e9943144f9da5916","login":"stas-sl","name":"Stas","avatar_url":"https://avatars-next.observableusercontent.com/7d59579aae90327bd1b21be4f9b238eb7bcb80a7678a195cfdc1f25b399d53b4","tier":"starter_2024"},"version":2274,"publish_version":null,"role":null,"authored":false,"authors":[{"type":"individual","id":"5fde748e49a846c1","login":"stas-sl","name":"Stas","avatar_url":"https://avatars-next.observableusercontent.com/7d59579aae90327bd1b21be4f9b238eb7bcb80a7678a195cfdc1f25b399d53b4","tier":null}],"members":[{"type":"individual","id":"5fde748e49a846c1","login":"stas-sl","name":"Stas","avatar_url":"https://avatars-next.observableusercontent.com/7d59579aae90327bd1b21be4f9b238eb7bcb80a7678a195cfdc1f25b399d53b4","tier":null,"role":"editor"}],"comments":[{"id":"831eb4d461bef37c","reply_to_id":null,"content":"Hey Stas, just wanted to thank you for sharing your great work. I got ton of insights and inspiration and I am going to use your clubhouse client and also learned from you to use it as a SQL cell. It's amazing! Thanks!","create_time":"$D2022-08-02T15:41:58.474Z","update_time":null,"resolved":true,"user":{"type":"individual","id":"35030f2775c5ade5","login":"oluckyman","name":"Ilyá Belsky","avatar_url":"https://avatars-next.observableusercontent.com/4fd43f90fde77934b5f4caa4453accf60bdeeb2cc01f813bfc0ad738c2b77958","tier":null}},{"id":"26a69a26d4433ea3","reply_to_id":"831eb4d461bef37c","content":"Glad, it was helpful to you 🙂","create_time":"$D2022-08-02T18:31:40.251Z","update_time":null,"resolved":true,"user":{"type":"individual","id":"5fde748e49a846c1","login":"stas-sl","name":"Stas","avatar_url":"https://avatars-next.observableusercontent.com/7d59579aae90327bd1b21be4f9b238eb7bcb80a7678a195cfdc1f25b399d53b4","tier":null}}],"custom_thumbnail":true,"trashed":false,"subscribed":null,"forks":3,"collection":null,"type":"notebook","body":{"title":"GitHub Issues Survival Analysis","cells":[{"id":0,"value":"# GitHub Issues Survival Analysis","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1524,"value":"$31","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1509,"value":"Here is how the data looks like:","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1502,"value":"select created_at, event_type, repo_name, number as issue, action, title, actor_login as actor, labels\nfrom github_events\nwhere event_type = 'IssuesEvent' and repo_name in (${repoFormatted})\norder by created_at desc\nlimit 100;","pinned":true,"hidden":false,"mode":"sql","author":"$undefined","summary":"$undefined","output":"","database":"var:client"},{"id":1552,"value":"To build the burndown chart we need to:\n 1. Combine 2 rows: _open_ and corresponding _close_ event in one row using ASOF JOIN to find closest records.\n 2. Repeat each row representing time interval, when issue was open from the previous result, as many times as there are days inside this interval using CROSS JOIN (cartesian product) with all dates.\n 3. Group by \"date\" **and** \"issue creation date\" to get the counts. It might be a good idea to use more coarse intervals for aggregation to not end up with a huge result set.\n\n_Note, sometimes you might get incomplete results because of repo being private in the early days, or if an issue was deleted instead of closed, nevertheless I believe the overall trend could be well visible._","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1587,"value":"In the textbox below you can enter your favorite repo name (case sensitive) to explore its issues. If at some point the repo changed its name/moved, you can provide multiple names separated by comma to merge all the issues together. ","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":29,"value":"viewof repo = Inputs.text({\n label: \"Repo\",\n submit: \"Go\",\n width: 450,\n value: \"ClickHouse/ClickHouse, yandex/ClickHouse\",\n datalist: [\n \"ClickHouse/ClickHouse, yandex/ClickHouse\",\n \"duckdb/duckdb, cwida/duckdb\",\n \"observablehq/plot\",\n \"observablehq/feedback\",\n \"vuejs/vue\",\n \"facebook/react\",\n \"pytorch/pytorch\",\n \"tensorflow/tensorflow\",\n \"googlecolab/colabtools\",\n \"bokeh/bokeh\",\n \"ecomfe/echarts, apache/incubator-echarts, apache/echarts\",\n \"vega/vega\",\n \"vega/vega-lite\",\n \"altair-viz/altair\",\n \"matplotlib/matplotlib\",\n \"mwaskom/seaborn\",\n \"plotly/plotly.js\",\n \"plotly/plotly.py\",\n \"pandas-dev/pandas\"\n ]\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":754,"value":"$32","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":2217,"value":"$33","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1366,"value":"viewof histOpts = Inputs.form({\n maxbins: Inputs.range([10, 100], {label: \"Max bins\", value: 20, step: 1}),\n ageFilter: Inputs.range([0, 1], {label: \"Max age (quantile)\", value: 0.95, step: 0.001}),\n state: Inputs.checkbox([\"open\", \"closed\"], {label: \"State\", value: [\"open\", \"closed\"]})\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1317,"value":"$34","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1632,"value":"A more fancy way to visualize how long it takes to resolve an issue, is to use Kaplan-Meier method, traditionally used in medical research to measure how long do people live after the onset of a disease, but it also works whenever you need to measure time to some event. It may sound like something complicated, if you've never heard about it before, but in fact it is not much harder to implement than a plain histogram. [This](https://allendowney.github.io/SurvivalAnalysisPython/02_kaplan_meier.html) introduction article helped me to understand the concept.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1754,"value":"viewof splitKM = Inputs.toggle({label: \"Split by year?\", value: true})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1640,"value":"$35","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1859,"value":"For convenience I also generated some burndown charts for my favorite repos (click on image to enlarge):","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1781,"value":"htl.html`

\n ${pregeneratedCharts.map(\n (x) =>\n htl.html.fragment`\n

${x.name}

x.img.width = x.img.width !== 250 ? 250 : width}}>${x.img}

`\n )}\n

`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1593,"value":"### Random findings: Facebook's React repo drama\nYou might have noticed that recently Facebook's React repo had big spike in number of issues, I was curious what that could be, and also to check correctness of my code, I looked into the data of that period and that's what I found. Looks like it was people's (or bot's) reaction to Facebook's reaction to the war in the Ukraine. Right now all these issues are deleted, and unfortunately deletion is not tracked in the events, so the chart shows more open issues than they exist in reality, but that's probably the least of a problem in this situation.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":1583,"value":"select created_at, event_type, repo_name, number as issue, action, title, actor_login as actor, labels\nfrom github_events\nwhere event_type = 'IssuesEvent' and repo_name = 'facebook/react' and toDate(created_at) = '2022-03-03'\norder by created_at\nlimit 100;","pinned":true,"hidden":false,"mode":"sql","author":"$undefined","summary":"$undefined","output":"","database":"var:client"},{"id":1464,"value":"## Appendix","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined"},{"id":3,"value":"import {client} from '@stas-sl/clickhouse-playground'","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":661,"value":"embed = require(\"vega-embed@6\")","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1453,"value":"mutable firstRun = true","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1494,"value":"repoFormatted = repo\n .split(\",\")\n .map((x) => `'${x.trim()}'`)\n .join(\",\");","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1272,"value":"$36","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1382,"value":"vegaConfig = ({\n renderer: \"svg\",\n actions: true,\n config: {\n axis: {\n titleFontSize: 14,\n labelFontSize: 13,\n titleFont: \"system-ui\",\n labelFont: \"system-ui\"\n },\n legend: {\n titleFontSize: 14,\n labelFontSize: 14,\n titleFont: \"system-ui\",\n labelFont: \"system-ui\"\n },\n title: {\n fontSize: 16,\n titleFont: \"system-ui\",\n labelFont: \"system-ui\"\n }\n }\n})","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"},{"id":1776,"value":"pregeneratedCharts = {\n let attachments = [\n FileAttachment(\"clickhouse.png\"),\n FileAttachment(\"duckdb.png\"),\n FileAttachment(\"observable-feedback.png\"),\n FileAttachment(\"observable-plot.png\"),\n FileAttachment(\"pytorch.png\"),\n FileAttachment(\"tensorflow.png\"),\n FileAttachment(\"bokeh.png\"),\n FileAttachment(\"matplotlib.png\"),\n FileAttachment(\"pandas.png\"),\n FileAttachment(\"plotly.js.png\"),\n FileAttachment(\"echarts.png\"),\n FileAttachment(\"seaborn.png\"),\n FileAttachment(\"vega-lite.png\"),\n FileAttachment(\"vega.png\"),\n FileAttachment(\"altair.png\"),\n FileAttachment(\"react.png\"),\n FileAttachment(\"vue.png\")\n ];\n const names = attachments.map(x => x.name.split('.')[0]);\n const imgs = await Promise.all(attachments.map(x => x.image({width: 250})))\n return _.zip(names, imgs).map(([name, img]) => ({name, img}));\n}","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined"}],"files":"$W37","stdlib":"1","titling":"auto"}} 30:"$2f:body"