<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Chenjing.Claire on Medium]]></title>
        <description><![CDATA[Stories by Chenjing.Claire on Medium]]></description>
        <link>https://medium.com/@chenjing.claire?source=rss-c0703fd48882------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*VIBuHJH4aPJJrodFRadHoQ.jpeg</url>
            <title>Stories by Chenjing.Claire on Medium</title>
            <link>https://medium.com/@chenjing.claire?source=rss-c0703fd48882------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 23:33:06 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@chenjing.claire/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[為什麼要花1.2萬元買一張女人.png? 紀錄第一個NFT收藏]]></title>
            <link>https://medium.com/@chenjing.claire/%E7%82%BA%E4%BB%80%E9%BA%BC%E8%A6%81%E8%8A%B11-2%E8%90%AC%E5%85%83%E8%B2%B7%E4%B8%80%E5%BC%B5%E5%A5%B3%E4%BA%BA-png-%E7%B4%80%E9%8C%84%E7%AC%AC%E4%B8%80%E5%80%8Bnft%E6%94%B6%E8%97%8F-37d9a949610d?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/37d9a949610d</guid>
            <category><![CDATA[nft]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[ethereum]]></category>
            <category><![CDATA[cryptocurrency]]></category>
            <category><![CDATA[blockchain]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Tue, 10 Aug 2021 07:33:25 GMT</pubDate>
            <atom:updated>2021-08-10T07:39:42.031Z</atom:updated>
            <content:encoded><![CDATA[<p><em>嘗試以消費者/收藏者（而非極客/技術者）的角度去理解NFT。</em></p><h4>我的NFT時序里程碑</h4><p>4/5 第一次聽到NFT這三個字，是跟<a href="https://www.theverge.com/2021/3/11/22325054/beeple-christies-nft-sale-cost-everydays-69-million">Beeple拍出6900萬的&lt;Everydays: The First 5000 Days&gt;</a>一起認識的，彷彿夢回2017年的幣圈，貴圈實在太瘋了。</p><p>6/24 看再多文章、聽再多podcasts，也不如下海自己鑄造一枚NFT，於是我在OpenSea上 list了一幅原創塗鴉 0.05 WETH起標，至今仍乏人問津，但花了一筆gas fee了解一枚NFT是如何誕生的、OpenSea跟其他平台有什麼差別，還算是值得。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*q6OrO2fHUJxEWfRJt6hmqA.png" /></figure><p>8/2 觀望了四個月，從NBA Top Shot, VeeFriends, CryptoPunks, 到Bored Ape Yacht Club，除了買不起還是買不起…終於等到一個買的起也認可的Collection，她是一幅名為”<a href="https://opensea.io/assets/0xe785e82358879f061bc3dcac6f0444462d4b5330/4877">Woman #4877</a>&quot;的 4000x4000 png檔，當時標價”0.175 ETH&quot;，換算成台幣是 12,635 NTD。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FarxCZ1dyTy0vMVPen5P6A.png" /><figcaption>我的第一枚NFT ”<a href="https://opensea.io/assets/0xe785e82358879f061bc3dcac6f0444462d4b5330/4877">Woman #4877</a>&quot;</figcaption></figure><h4>關於 NFT Collection — “World of Women”</h4><p>如果說我對NFT有任何信念的話，八成是這位心靈雞湯大叔 <a href="https://www.instagram.com/garyvee/">GaryVee</a> 給灌的。8/1 GaryVee 發了一則裸上身(?)的自拍視頻，介紹了<a href="https://worldofwomen.art/">World of Women</a> 這個NFT Collection，他說「如果你的預算不高但想多了解NFT ，那麼可以考慮這個project，從著手購買ETH、設定MetaMask、在OpenSea上購買NFT，實際的操作整個流程。」That’s me! 雖然買不起他的<a href="https://veefriends.com/">VeeFriends</a>(現在最便宜的一個要價台幣50萬+)，總該買的起他第一次開金口報的明牌。於是我開始研究起”女人的世界”。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I8Lw3cNmAeTLsGxPIM82SQ.png" /><figcaption><a href="https://www.instagram.com/p/CSCO3augnEc/">https://www.instagram.com/p/CSCO3augnEc/</a></figcaption></figure><blockquote>World of Women aims to bring in more diversity and inclusivity to nfts, which we believe is very much missing in this male dominated space.</blockquote><p>World of Women背後的藝術家是<strong>Yam Karkai</strong>，「她一直以來的創作都致力於通過讓女性成為她作品的中心來突出並賦予女性權力。」</p><p>World of Women是10,000名隨機生成的女人，各自有獨特的特徵，如背景圖、髮型、嘴唇、膚色、配件、衣服、面部特徵等等。只要成為Woman NFT owner就可以享有50%的商業版稅，也就是如果未來這些女人作為商業用途賺了錢，那麼所有女人的NFT owner都可以獲得50%的版稅。另外，其中一部分的”特殊女人”屬於三個俱樂部：ROYALTIES CLUB的成員每個月可以平分所有二手交易的2%版稅，INVESTORS CLUB的成員每個月可以平分WoW基金的50%盈利，CURATORS CLUB的成員每個月可以選一枚NFT藝術作品讓WoW基金購買，以支持他們最喜歡的加密藝術家。</p><p>除了讓所有NFT擁有者可以互利互惠之外，WoW也把他們的理念投入且回饋到真實世界中，其中2.5%的盈利會捐出用於支持公益組織，包含 <a href="https://shesthefirst.org/">SHE’S THE FIRST</a> 和<a href="https://tooyoungtowed.org/">TOO YOUNG TO WED</a>，以及15%的盈利將再投入加密藝術圈。</p><p>了解到這裡，我才發現之前認識的NFT項目除了價格難以高攀外，也幾乎都是男性導向的（e.g. NBA Top Shot, Sorare，VeeFriends, CryptoPunks, Bored Ape Yacht Club, etc.)，直到World of Women的出現讓NFT Space多了些女性的共鳴。</p><p>於是終於在觀望四個月後，因為相信GaryVee的眼光、支持WoW的理念、想跳進NFT世界一探究竟，買了我的第一枚NFT — ”<a href="https://opensea.io/assets/0xe785e82358879f061bc3dcac6f0444462d4b5330/4877">Woman #4877</a>&quot;。</p><figure><img alt="08/02/2021 World of Women on OpenSea截圖：2.6K owners, 0.18 floor price, 1.0K volume trade" src="https://cdn-images-1.medium.com/max/1024/1*jmDli2i1wdvPA4nBot3BtA.png" /><figcaption>截圖於 08/02/2021</figcaption></figure><h4>為什麼要花1.2萬元買一張女人.png?</h4><p>先撇除NFT不知所謂何物的概念，以普遍的收藏品(collectibles)來說，我們購買它不外乎以下兩個原因：</p><ol><li>擁有它並想跟其他人展示：例如，限量球鞋、奢侈品、遊戲王卡</li><li>相信它的增值空間：例如，年輕藝術家的作品</li></ol><p>但是以上的舉例都是觸碰得到的收藏品，為什麼我們要花錢買一個數位/虛擬檔案呢？不是所有人都可以複製貼上同一個圖檔嗎？</p><p>就像H&amp;M 之於 Hermès，前者因為是大量生產、沒有炫耀價值也不保值，後者因為稀缺且有品牌價值，消費者在乎他的真偽。同理NFT讓一個可以被輕易複製的數位檔案有了稀缺性和可鑑別真偽的能力，讓消費者可以以收藏的心態購入它背後的價值，如社會地位和增值潛力。</p><p>1.2萬元買的不只是這張”女人 #4877.png”，還有精神層面因支持WoW做公益而得到富足、展示我是Women NFT的一員而感到獨特、相信未來會有更多的人加入而使我的這枚NFT得到增值。</p><h4>如果NFT要再破圈，我認為還有三個關鍵…</h4><p>2021年初應該算是NFT的第一次破圈，但距離成為主流應用還需要解決沒地展示、缺乏實用性和使用門檻的問題。</p><p>最關鍵的一步會是 Facebook, Instagram, TikTok, WhatsApp等主流的社交App開始讓用戶在個人頁中展示所擁有的NFTs，那麼NFTs的可展示性以及炫耀和社群價值將會大幅提升。如果主流App不這麼做很可能會有新的Social App因此崛起並搶佔他們的龍頭地位。</p><p>再來是更多NFT衍生的產品、服務和社群來解決缺乏實用性的問題，例如現在有些NFT是包含會員獨家內容、福利、或VIP活動，但都是依靠對創作者的信任完成交易的，缺乏與NFT串連的平台去系統化完成。當這些功能被產品化後可以為很多NFT提升額外的價值和服務，這時對於品牌忠誠度較一般、更關注實用性的大眾消費者來說才有足夠的動機去購買這類的NFT。</p><p>Last but not least, NFT的使用和學習門檻必須降低。從了解虛擬貨幣、購買ETH、註冊一個加密錢包、導入ETH到加密錢包、了解NFT、到購買NFT，整個學習過程彷彿攀登珠穆朗瑪峰(?)，必須讓他跟在蝦皮/Amazon上購物一樣容易，才能讓一般消費者輕易抵達。</p><p>最後，「投資一定有風險，Crypto/NFT投資有賺有賠，申購前應詳閱公開說明書。」目前大部分的NFT Collections恐怕多是投機者炒作出來的泡沫，在繳學費的過程中，祝我們都保持理性和客觀。</p><p>如果你也相信NFT的未來、想要一起成為破圈的先行者，歡迎加入我們的新創團隊(<a href="https://www.victor.ly/">victor.ly</a>)，打造有意義且有價值的產品。</p><p>聯繫信箱：claire@victor.ly</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=37d9a949610d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Quarter-life crisis?]]></title>
            <link>https://medium.com/@chenjing.claire/quarter-life-crisis-e3fd6e51d6f1?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/e3fd6e51d6f1</guid>
            <category><![CDATA[職涯]]></category>
            <category><![CDATA[互連網]]></category>
            <category><![CDATA[青年危機]]></category>
            <category><![CDATA[工作]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Sat, 29 Aug 2020 15:08:29 GMT</pubDate>
            <atom:updated>2020-08-29T15:08:29.461Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*m1AnAPKm_CcZEVe0.jpg" /></figure><p>Hey future me,</p><p>剛過了25歲的生日，沒太感受到同齡人常提起的 quarter-life crisis (青年危機)，可能是因為2020太魔幻，以致於不奢望能控制未知，只要能平安健康的生活著、甚至還做著一份能帶來成長的工作，就已經非常滿足了，即使工作環境本身也是個被時代洪流牽動的漩渦。</p><p>最近的體會是，只要有夠強大的救生衣，即使洪水再湍急，也可以相對平穩地飄在浪上，靜待著潮流把你帶向哪，因為有足夠的信心，最終可以安穩的抵達彼岸，還有能力再站起來。</p><p>沒有感受到 quarter-life crisis 也有可能是因為身邊朋友同事們，大多都還生活的跟20出頭時一樣，週末相約吃Brunch、聽Live，讓我對30歲的想像，不再像爸媽那輩的 quarter-life — 有車、有房、有子x3 — 那麼 intense😂</p><p>老實說，駕駛在高速互連網賽道，很難避免年齡焦慮。我們的 CEO 才37歲，他29歲創的這家公司，估值已經750億美元了，而我們的中層Leadership 大多都不到30歲，坊間盛傳著35歲，沒當上 leader 就會被互連網行業淘汰。Forbes 雜誌也老愛宣傳30 under 30/ 40 under 40，好像不在某個年齡前，改變世界、實現財富自由，往後餘生就抬不起頭似的🙃</p><p>可是，人生是個無限遊戲啊。不像足球或下棋那樣的有限遊戲，有已知的玩家、設定好的規則跟終點線，能容易的分辨出輸家跟贏家。人生中的玩家來來往往，規則瞬息萬變，唯一的終點線是生命的盡頭，也因此不會有真正的贏家和輸家。有限遊戲才需要爭輸贏，而人生這場無限遊戲，我們只需要盡可能的留在賽局裡，每天比昨天還更好。</p><p>最後銘記，</p><blockquote>只要活的夠久，年齡就不是問題。</blockquote><p>望你十年後，跟現在一樣不被年齡所困🥳</p><p>Best,</p><p>2020.08.29</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Ftye525dkfi8%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dtye525dkfi8&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2Ftye525dkfi8%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/6813b17a8a0026bdb46e66879646d579/href">https://medium.com/media/6813b17a8a0026bdb46e66879646d579/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e3fd6e51d6f1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[PM如何做好信息路由＆面試中看重的特質 | PM每週Pep Talk]]></title>
            <link>https://medium.com/@chenjing.claire/pm%E5%A6%82%E4%BD%95%E5%81%9A%E5%A5%BD%E4%BF%A1%E6%81%AF%E8%B7%AF%E7%94%B1-%E9%9D%A2%E8%A9%A6%E4%B8%AD%E7%9C%8B%E9%87%8D%E7%9A%84%E7%89%B9%E8%B3%AA-pm%E6%AF%8F%E9%80%B1pep-talk-8d398802fb5d?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/8d398802fb5d</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[product-manager-interview]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[product-management]]></category>
            <category><![CDATA[agile-development]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Mon, 23 Mar 2020 18:28:15 GMT</pubDate>
            <atom:updated>2020-03-23T18:28:15.702Z</atom:updated>
            <content:encoded><![CDATA[<p><em>【PM每週Pep Talk】記錄了我們每週週會上的吐槽、學習、見解、近期思考、等無邊界的話題。稱之為Pep Talk，是因為我們藉由每週1.5小時的抒發，幫助彼此回到崗位上繼續更穩定＆高效的工作 :P</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5R2aRXMhdcAGBF5_b_-vGA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ptRc6W-G605Z6ns3KorPNg.png" /></figure><h3>Topic 1: PM如何做好信息路由</h3><h4><strong>做信息路由对PM的好处</strong></h4><ul><li>PM之所以被交付很多責任與決定權，是因為這個角色能接受到最全面的信息，而作為一個好的信息路由，能幫助我們自己釐清和拼凑各方信息，更全面地理解跨部門、跨業務的合作脉络和信息全貌</li><li>做一個可靠的信息同步方，能去得團隊的信任與信賴，也有”穩定軍心”的作用，讓研發安心的開發、讓設計安心的畫圖，因為他們相信你會把其他事情搞定</li><li>當PM把信息同步的工作做的完善時，也能同時提升工作效率，因為能減少一天當中各方需要你解答的次數</li><li>有時需求/項目無法推展下去，不是因為不可行，而是沒有足夠的信息和對應人來補充背景和決策，而做好信息路由就能避免自己成為blocker</li></ul><h4><strong>好的信息同步方式</strong></h4><ul><li>向所有人同步完整的背景、清楚介紹緣由和要解決什麼問題，表明需要各方提供哪些帮助或額外信息，確保大家on the same page</li><li>客觀、清晰的陳述信息（是什麼、為什麼、決定是什麼），不添加自己的主觀判断</li><li>PM將自己負責的重點方向介紹、常見的問題、以及常合作的對接人名單”文檔化”，就能又一句話”看我文檔”取代一次又一次的人工答錄機</li><li>做筆記、會議紀要：Inspired by the post &lt;<a href="https://medium.com/swlh/i-learned-at-least-6-things-in-my-first-week-as-a-product-manager-2515a67a141b"><em>I learned at least 6 things in my first week as a Product Manager</em></a>&gt;</li></ul><blockquote>Speaking of<strong> </strong>building trust — which is kind of important in a role all about influencing an outcome <strong>without any explicit authority</strong> — a really counterintuitive way to do that is to do a thing that a lot of PMs may feel is the most basic, junior, or degrading part of the role: take great notes.<br>Take notes in team meetings, in 1:1s, in meetings with partners and stakeholders. Email the notes to the relevant parties, ensuring to highlight key decisions, action items, or key context.<br>Capturing thoughts, clarifying them, summarizing them, and turning them into action is an opportunity to <strong>show the people you are working with that their thoughts matter</strong> and can be the difference between idle chatter and a meeting that pushes the project forward. Doing this consistently will <strong>build trust and credibility</strong>, giving you a much better shot at having your own ideas taken seriously.</blockquote><h4>不好的信息同步方式</h4><ul><li>沒有搞明白對方的問題和訴求，給予錯誤的判斷或指引到錯誤的方向</li><li>當被問到不清楚的業務時，没有借助自己的connection帮忙牽線，成為了事情的blocker</li><li>沒有把所有應該被同步到、有決策權的各方人拉齊再同步，導致決策有偏差或是需要反覆介紹和溝通</li></ul><h3>Topic 2: 面試PM中看重的特質</h3><ul><li>不僅展現出對工作的熱情，還有對於生活的熱情，講到熱愛的事眼神會散發出光</li><li>有邏輯的語言組織、簡短到位的表達，避免囉嗦、空洞的長篇大論</li><li>能獨立思考，不被假設問題帶跑，從問題根源出發開始分析、以及用客觀事實來論證，而不是主觀直覺判定，或讀網上文章照本宣科</li><li>能迅速發現自己思路的問題（反應快），承認錯誤並調整（謙遜），有反思、從錯誤中學習、和主動尋求外界幫助的能力</li><li>抗壓能力強，能在連環逼問中自信且從容的應答</li><li>同理心，能站在別人的角度、用對方懂的語言去溝通</li><li>好奇心强，對不了解的事有主動學習的意願，相反的如果連應聘的公司和產品都沒有研究，不僅沒有好奇心，也不夠重視這個機會</li><li>展現出對於這個機會的期待與非我不可的決心（例如最基本的守時、適宜的服裝、良好的網速、安靜的環境）</li></ul><p><em>*Credit to TikTok PM squad.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8d398802fb5d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[記錄Coronavirus的觀察和反省]]></title>
            <link>https://medium.com/@chenjing.claire/%E8%A8%98%E9%8C%84coronavirus%E7%9A%84%E8%A7%80%E5%AF%9F%E5%92%8C%E5%8F%8D%E7%9C%81-85bfa7130ee?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/85bfa7130ee</guid>
            <category><![CDATA[coronavirus]]></category>
            <category><![CDATA[2020]]></category>
            <category><![CDATA[opinion]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Sun, 22 Mar 2020 15:12:49 GMT</pubDate>
            <atom:updated>2020-03-22T15:12:49.422Z</atom:updated>
            <content:encoded><![CDATA[<p>「把自己抽離去看待正在進行中的一切。」那天週會上，我們討論到感覺迷惘/壓抑/負面的時候，怎麼去調整心態。也許人生的本質就是體驗，好的壞的都是一番體驗，唯有這麼去理解悲劇，才不致於深陷其中無法脫身。</p><iframe src="https://www.instagram.com/p/B9t4cssnxOk/embed/?cr=1&amp;amp;rd=https%3A%2F%2Fwww.instagram.com" width="658" height="882" frameborder="0" scrolling="no"><a href="https://medium.com/media/2497280ef7dff0c64dedb9e86e3aaf8a/href">https://medium.com/media/2497280ef7dff0c64dedb9e86e3aaf8a/href</a></iframe><p>寫在3/22，此時此刻全球被Coronavirus肆虐，除了中國本地確診已經零新增，我生活在的上海也已經慢慢復甦了，展覽、網紅店開始排起人龍。</p><p>回想起2019跨新年時，我短暫地回到台灣，看到一則小道新聞&lt;武漢出現多起不明肺炎…疑似SARS…&gt;我在爸媽的辦公室裡(旅行社)念出來這個聳動的標題，同事們忙著處理手上多筆訂單，頭也沒抬地說「不要吧～像當年SARS那樣就沒生意做了…」，畢竟新聞出處不是官媒，我們都沒放在心上。</p><p>跨年後回到上海開始上班，風平浪靜的直到1/22除夕前一週，事情變得愈來愈嚴重，大家開始戴起口罩，媽媽是醫務人員的同事告誡我們真實情況可能比新聞說的、我們想的還要嚴重很多。除夕前兩天，我按照計畫回台灣過年，臨走前開玩笑的跟同事說，「我們可能一個月後再見囉！」，真沒想到我們後來整整兩個月後才在辦公室相見。</p><p>這兩個月中，見證了武漢的爆發、擴散、封城，在慢慢到清零，從恐慌、混亂、憤怒，到平靜、康復。與此同時，地球的另一端重蹈覆轍的再一次經歷同樣的劇情。很幸運的，我和家人都不在風暴的中心，才能相對抽離的去看待這魔幻的2020開幕…</p><figure><img alt="2019年底買了這本經濟學人2020預測，現在看起來就像本童話故事&lt;沒有Coronavirus的2020應該是什麼樣子&gt;" src="https://cdn-images-1.medium.com/max/1024/0*rtt3SCPnOmllCDpt.jpg" /><figcaption>2019年底買了這本經濟學人2020預測，現在看起來就像本童話故事&lt;沒有Coronavirus的2020應該是什麼樣子&gt;</figcaption></figure><h4>保持善良和理性</h4><p>病毒對大多數人來說，可能帶來的只是短期的不便，幾個月後生活又會回歸正常，但難以癒合的是非理性和仇恨帶來的對立。當川普執意將種族意識扣上病毒時，你知道他的算盤。</p><h4>全球化</h4><p>記得有個道理是這樣的，「一串鎖鏈的強度取決於最弱的那一環」。一個過度依賴貿易、強調市場經濟的現代社會，在面臨這樣的全球瘟疫時，變得非常脆弱，只要一個環節斷了，整個體系就有崩壞的風險。無止盡地追求利益最大化，是我們需要的嗎？</p><h4>效益主義 vs 正義論</h4><p>效益主義的思想是「將行為帶來的總體快樂減去痛苦，能將效益達到最大化，就是道德的」，而正義論的思想是「最大化最小數，也就是即使整體平均效益下降，也該最大化最弱勢的權益」。</p><p>例如英國的全民免疫政策、美國有限的檢測試劑需要靠錢和關係才能取得，忽視弱勢的老人們和底層人民的權益，都不是正義論所提倡的。例如政府下令居委處強制全民居家隔離，且實施實名簽到追蹤，導致整體效益下降、損害權益，就不是效益主義所支持的，而這樣的措施卻可以保障弱勢族群的權益。</p><p>多點思考、多點同理心，stay safe, stay healthy.</p><p>最後我建議多看看TikTok，你會發現世界很幽默。</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.tiktok.com%2Fembed%2F6805421968276409606&amp;display_name=tiktok&amp;url=https%3A%2F%2Fwww.tiktok.com%2F%40mikhuul%2Fvideo%2F6805421968276409606&amp;image=http%3A%2F%2Fp16-va-default.akamaized.net%2Fimg%2Ftos-maliva-p-0068%2F6c3c3108ec864a419612336ef2285acc_1584510788%7Etplv-tiktok-play.image&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=tiktok" width="340" height="700" frameborder="0" scrolling="no"><a href="https://medium.com/media/94049f88c810f8735b67ddc85801af70/href">https://medium.com/media/94049f88c810f8735b67ddc85801af70/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.tiktok.com%2Fembed%2F6804228408701095174&amp;display_name=tiktok&amp;url=https%3A%2F%2Fwww.tiktok.com%2F%40ashleytisdale%2Fvideo%2F6804228408701095174&amp;image=http%3A%2F%2Fp16-va-default.akamaized.net%2Fimg%2Ftos-maliva-p-0068%2Fed3cd51adea7e9ac053d240a090535a7%7Etplv-tiktok-play.image&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=tiktok" width="340" height="700" frameborder="0" scrolling="no"><a href="https://medium.com/media/29e9f30de2ec95474317bc2ccdb2f73a/href">https://medium.com/media/29e9f30de2ec95474317bc2ccdb2f73a/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.tiktok.com%2Fembed%2F6805716271477312773&amp;display_name=tiktok&amp;url=https%3A%2F%2Fwww.tiktok.com%2F%40houseofhighlights%2Fvideo%2F6805716271477312773&amp;image=http%3A%2F%2Fp16-va-default.akamaized.net%2Fimg%2Ftos-maliva-p-0068%2F1ca3e73df8725bd4932a56e0977f2e74%7Etplv-tiktok-play.image&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=tiktok" width="340" height="700" frameborder="0" scrolling="no"><a href="https://medium.com/media/6183e04dfa77e010e422fbdf572bd7e4/href">https://medium.com/media/6183e04dfa77e010e422fbdf572bd7e4/href</a></iframe><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.tiktok.com%2Fembed%2F6806608302705855746&amp;display_name=tiktok&amp;url=https%3A%2F%2Fwww.tiktok.com%2F%40callmegrand%2Fvideo%2F6806608302705855746&amp;image=http%3A%2F%2Fp16-sg-default.akamaized.net%2Fimg%2Fv0201%2F68ec5cbf021477bf8a3469bf5d3fa46b%7Etplv-tiktok-play.image&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=tiktok" width="340" height="700" frameborder="0" scrolling="no"><a href="https://medium.com/media/5d13e16ba722aad608cc9fc954a95e5d/href">https://medium.com/media/5d13e16ba722aad608cc9fc954a95e5d/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=85bfa7130ee" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[細思極恐的人生選擇 Go take the road less traveled by]]></title>
            <link>https://medium.com/@chenjing.claire/%E7%B4%B0%E6%80%9D%E6%A5%B5%E6%81%90%E7%9A%84%E4%BA%BA%E7%94%9F%E9%81%B8%E6%93%87-go-take-the-road-less-traveled-by-56beef6ca4e7?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/56beef6ca4e7</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[careers]]></category>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[startup]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Sat, 21 Mar 2020 06:38:51 GMT</pubDate>
            <atom:updated>2020-03-23T18:22:43.713Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>Two roads diverged in a yellow wood,<br>And sorry I could not travel both<br>And be one traveler, long I stood<br>And looked down one as far as I could<br>To where it bent in the undergrowth</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*av9rML47bpmwmCl2U4HwUQ.png" /></figure><p>如果我高二那年按部就班的繼續準備學測，那麼我可能會在某國立大學裡庸庸碌碌的荒廢四年青春，不曾想過去到小時候看的那部電影&lt;金髮尤物Legally Blonde&gt;的拍攝場景裡念書。</p><p>如果我大四那年決定加入任何一家灣區互連網公司，那麼我可能不會研究起中國的互連網產業、也不會決定搬到中國工作。</p><p>如果我當時妥協到杭州某家公司工作，那麼我可能就不會加入一個後來成為繼Instagram後最火的產品。</p><p>如果我走上了另一條路，那我今天會在哪裡、跟什麼樣的人、做什麼樣的事？</p><blockquote>I shall be telling this with a sigh<br>Somewhere ages and ages hence:<br>Two roads diverged in a wood, and I — <br>I took the one less traveled by,<br>And that has made all the difference.</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*vU75UQVa1y--dVMn.gif" /><figcaption>From &lt;Legally Blonde&gt;, one of my all-time-fav movie</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*Uyht9sTUY4kfC0QD" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=56beef6ca4e7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[從夢想在矽谷工作，到落腳上海互聯網公司的這條路]]></title>
            <link>https://medium.com/@chenjing.claire/%E5%BE%9E%E5%A4%A2%E6%83%B3%E5%9C%A8%E7%9F%BD%E8%B0%B7%E5%B7%A5%E4%BD%9C-%E5%88%B0%E8%90%BD%E8%85%B3%E4%B8%8A%E6%B5%B7%E4%BA%92%E8%81%AF%E7%B6%B2%E5%85%AC%E5%8F%B8%E7%9A%84%E9%80%99%E6%A2%9D%E8%B7%AF-7d7044bcd94?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/7d7044bcd94</guid>
            <category><![CDATA[careers]]></category>
            <category><![CDATA[shanghai]]></category>
            <category><![CDATA[internet]]></category>
            <category><![CDATA[silicon-valley]]></category>
            <category><![CDATA[jobs]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Sun, 26 Aug 2018 14:08:48 GMT</pubDate>
            <atom:updated>2018-08-26T14:08:48.361Z</atom:updated>
            <content:encoded><![CDATA[<p>*<a href="https://crossing.cw.com.tw/blogTopic.action?id=534&amp;nid=10400">原文刊登於換日線</a></p><h3>— — — 決定「裸歸」篇</h3><p>「在 Airbnb 做數據分析、月入 20 萬（新台幣）以上、開 Tesla model 3 、住在舊金山的公寓⋯⋯」這是我從大一到大三，對未來畢業後的自己，始終如一，具體（而天真）的藍圖。</p><p>結果現在的我，住在上海黃埔區的一間老工房，每天走路上班，下班後偶爾騎騎共享單車、逛逛夜上海。但我非常快樂，也很慶幸。</p><p>而這一路以來的轉折，便是我在接下來的系列文章中，希望能分享給大家的經驗：</p><p><strong>自從四年前的大一，當我一來到美國加州後，就開始努力地為畢業後加入矽谷巨頭科技公司的夢想而努力著：</strong>不管是瞄準當下興起的職業 — — 數據分析師，而決定從經濟系轉到應用統計系；努力找實習、兼任新創公司的品牌大使；跟朋友們創立學校第一屆大型黑客松（<a href="https://en.wikipedia.org/wiki/Hackathon">Hackathon</a>）；到處參加相關的演講跟競賽⋯⋯無不是為了在四年畢業後，能順利地找到一份夢想中的工作。</p><p>但儘管一步一步地努力往這樣的方向爬，到了大三為止，終究還是觸不到巨頭公司的一點兒邊。要繼續「非矽谷巨頭不進」地努力到簽證到期前的最後一刻？還是同步嘗試其他的選項？我開始重新審視自己的條件與考量重點。</p><p><strong>重新思考、分析自己的「夢想工作」</strong></p><p>抱著同樣的夢想努力三年多後，這是我第一次靜下心想，留在美國工作，對我來說真的是最好的選擇嗎？</p><p>大部分選擇留在美國工作的兩岸留學生，主要考量的重點有以下四個：</p><p><strong>1. 薪資</strong></p><p>薪資大概佔了決定因素的一半以上。同樣是軟體工程師，拿加州灣區跟北上廣深（北京、上海、廣州、深圳）的互聯網公司比，<strong>應屆畢業的起薪可以差到將近三倍。</strong>就算扣掉生活開銷的差距，同樣工作一年，灣區公司還是硬生生地多了整整一年的收入。如果是行銷、營運或設計類的職位，那差距就更大了。</p><p><strong>2. Work-life balance</strong></p><p>在美國，準點下班、週休二日、十幾二十天的年假都是「標準配備」。但不管在大陸還是台灣，平日加班、假日「 Line 上加班」、僅僅五天左右的年假⋯⋯才是常態。<strong>對已經習慣美國文化，重視自由與人權的人來說，可能很難重新去適應這樣的高壓生活。</strong></p><p><strong>3. 生活品質</strong></p><p>相對亞洲，美國西岸大城市空氣好、活動空間寬敞、建築整齊乾淨、人民素質也相對高，生活起來愜意又舒適，光是曬個太陽就有種「人生如此美好」的幸福感。回到亞洲，則可能要面臨天天擠捷運通勤、社會氣氛浮躁、霧霾籠罩、走兩步路就流汗的窘境。</p><p><strong>4. 光環</strong></p><p>最後一個，是大部分人只放在心底的考量：光環。只要說起兒女在美國工作，父母的臉上不免揚起得意的笑容 — — 不論他們在美國做什麼樣的工作、過什麼樣的生活。「來來來，來台大！去去去，去美國！」這種路徑給人「出人頭地」的印象，還深深地留在我們的觀念裡。</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/599/1*jnN7jhfJzTVHp_bSWWMfOg.png" /><figcaption>Venice, California. from <a href="https://www.instagram.com/chenjing.t.w/">@chenjing.t.w Instagram</a></figcaption></figure><p><strong>再一次定位自己的職涯路</strong></p><p>在關上所有聲音、好好的思考後，我發現上述的「Work-life balance」跟「生活品質」，對我來說都有很大的「交易空間」：譬如我願意用緊湊的生活步調，去交換「能在端午節吃上粽子、秋天吃上大閘蟹、過年吃上年夜飯」；我願意用擁擠的生活空間，去交換隨叫隨到的外賣服務（是的，我的鄉愁都與食物有關）；我願意用一切相對的辛勞與不適，去換來家的感受。</p><p>至於「薪資」的部分，如果不是太著急要在「幾歲前要賺到幾桶金」，我想應該可以把眼光放遠一點，看十年後我想在哪裡、做到什麼樣的職位 — — <strong>對我個人而言，雖然夢想畢業後在矽谷工作，但最終回到亞洲發展，是一直以來不變的目標。</strong></p><p>因為除了「鄉愁」的因素，在矽谷科技巨頭如果想做到管理階級，老實說不管公司思想有多開放公平，按照目前的圈內文化，絕大多數身為「外國人」的「亞洲」「女性」，還是會面臨到玻璃天花板 — — 反之，回到亞洲擁有留學背景，仍常是升遷的一大助力。</p><p><strong>於是說到底，我當初想留在矽谷的夢想，還是與「光環」有關：</strong>想證明自己能在競爭最激烈的人才市場搶下一個位子，以及證明大學學費的投資報酬率是正的。</p><p>雖然回到亞洲，一樣有機會證明這些事，只是時間會拉的更長，我願意等嗎？</p><p>為了粽子、大閘蟹跟年夜飯，我想我是願意的。於是我開始研究起大陸的互聯網公司 — — 在這之前，我對它們其實是接近一無所知的。什麼是 BAT （註二）、什麼是 TMD （註三）、什麼是秋招春招⋯⋯等等，都與美國是兩個完全不同生態系的市場。</p><p><strong>於是，在下定決心裸歸後，我便展開清晨四點起床參加線上筆試、白天上課、晚上十點後進行一連串視訊面試的日子⋯⋯。</strong></p><p>註：「裸歸」，意思是沒有工作經驗的海歸。<br>註二：BAT，是百度（Baidu）、阿里巴巴（Alibaba）、騰訊（Tencent）三家中國網路巨頭公司的合稱。<br>註三：TMD，是今日頭條（Toutiao）、美團點評（Meituan）、滴滴出行（DiDi）三家互聯網產業中的新興「獨角獸」企業合稱。</p><h3>— — — 說服自己和面試官篇</h3><p><strong>最常被問到的面試問題：為什麼要來中國？</strong></p><p>「我看你的履历非常优秀，美国留学、经历丰富，那为什么会选择来大陆工作？」</p><p>這是線上面試時，大多數面試官開頭會問我的第一個問題。我總不能說：「因為我在美國恐怕進不了巨頭公司，所以想試試大陸的獨角獸公司⋯⋯」吧，<strong>於是我認真地想想，除了「推力」之外，中國大陸的職場，給了我什麼「拉力」呢？</strong></p><p>上一篇文章中說到，大部分選擇留在美國工作的兩岸留學生，主要考量有「薪資」、「work-life balance」、「生活品質」，還有「光環」等。<strong>但其實還有一個沒提到的重要考量，是「產業發展」：</strong></p><p>以互聯網產業來說，中國市場的特色，除了龐大的數據量外，<strong>目前相對歐美等國來說，約束企業對數據應用的法規也「較為寬鬆」 — — 這對互聯網公司來說，是個很重要的資產，也是未來發展的關鍵因素。</strong></p><p>再加上中國市場大、人才多，目前國家更把科技產業、尤其互聯網產業當「大戰略」來執行，因此發展起來，彷彿是以「光速」進行的。</p><p>拿自動駕駛來說，「中國計劃在 2020 年前實現智能汽車數量，至少佔到全國汽車總量的一半，並且希望實現在 90% 的城市和道路上，使用自動駕駛汽車」 — — 這段話，引自中國<a href="http://www.ndrc.gov.cn/">發改委</a>提出的政策草案。我想大家應該可以想見，「在中國」若此案正式通過，可能很快就會立刻被「排除萬難、齊心執行」；但「在美國」，則可能到了 2020 年，國會還在討論自動駕駛發生事故責任歸咎的法律與道德問題，而因此遲遲不能推動自動駕駛合法上路。</p><p><strong>以上種種「環境因素」，讓我相信在可見的未來，至少在互聯網產業裡，中國企業有機會、也能夠急起直追美國企業的規模。</strong></p><p><strong>正在逐漸消失的「優勢」</strong></p><p>除了產業發展外，還有考慮到自身的職涯規劃：如果做數據分析師或工程師，這類偏向「純技術」的職位，那不可否認的，留在美國企業，能學得更多更精。但我對自己的了解與規劃，是往市場與產品端發展，不是做純技術。</p><p>於是在觀察了兩地的文化和產品設計的迥異後，我了解到<strong>「在美國積累的產品認知與市場敏銳度，是無法挪用至亞洲市場的，甚至有可能成為絆腳石」。</strong>因此盡早回到亞洲接軌，對我個人的長期職涯規劃（回到亞洲市場，請見上篇文章）來說，也許是個更好的安排。</p><p><strong>其實，還有個我沒跟面試官坦白的理由：因為海外學歷正漸漸貶值。</strong></p><p>特別是歷經半年的求職過程，我的感受更是深 — — 除了上面提到「外國經驗無法挪用到中國市場」的原因外，還有「越來越多的中國留學生回到中國工作」。</p><p><strong>今年，中國市場的「海歸」比例創了歷年紀錄，達到了 80% ，以後也只會更高 — — 當海外學歷的供給上漲後，需求甚至降低，可能再晚個幾年，就沒有優勢了。</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/598/1*AX0E9-tYktZzbyTuAiYGbw.png" /><figcaption>上海外灘街景 from <a href="https://www.instagram.com/p/Blb0UFZF_BY/?taken-by=chenjing.t.w">@chenjing.t.w Instagram</a></figcaption></figure><p><strong>為什麼選擇「上海」？為什麼選擇「有大企業撐腰的新創公司」？</strong></p><p>「为什么选择落脚上海？（而不是北、深、杭州或香港），为什么选择我们（ＯＯ企业）？」這是許多面試官的下一個問題。</p><p><strong>其實在我最後決定落腳上海之前，對亞洲不同的大城市，是沒有太大的主觀好惡的 — — </strong>不管是北京、杭州、深圳、香港、甚至是新加坡，我都有投履歷 — — <strong>因為除了我的故鄉台中以外，對我來說哪裡都是異鄉。</strong></p><p><strong>但當我收到了杭州和北京的 Offer ，並趁著寒假陸續造訪了這兩個城市和兩家公司的工作地點後，我發現 — — 我錯了，「異鄉」還是有差別的。</strong>我毅然決然重新開始投起第二波的履歷，單單瞄準上海的工作機會。</p><p>理由很簡單：天氣、文化、跟城市氛圍，都是我個人的決定因素 — — 兩年前來上海時，就喜歡上他的新舊融合，與東西文化合併的城市面貌。</p><p>至於選擇ＯＯ公司（其實不只一家），是看上它是在風口浪尖的領域，充滿挑戰與機會，工作過程中累積到的能力與經驗，都是將來對個人發展最有利的資本。<strong>它的公司規模與背景，我認為對應屆生來說也是最理想的：有大公司撐腰的中小型新創 — — 既穩定有組織，又能讓新人承擔很多責任，不只做一個小螺絲。</strong></p><p><strong>在說服別人的過程中，更明白自己真正要的</strong></p><p>在這半年說服眾多面試官的過程中，我漸漸更明白，自己真正想要什麼，也完全地說服了自己，選擇一條更適合的路。</p><p><strong>對比一下當時的</strong><a href="https://crossing.cw.com.tw/blogTopic.action?id=534&amp;nid=10400"><strong>夢想圖像</strong></a><strong>：</strong>我現在雖然沒有在矽谷的巨頭科技公司上班，但也到了一家佔據 App store 全球下載排行榜第一的公司 — — 能加入這樣像是 Instagram&amp;Snapchat 量級的團隊，是我那時留在美國搆也搆不到的機會。</p><p>雖然沒有開 Tesla model 3 上班，但能在晚上十點後放心的上街騎騎單車遊夜上海，是我在加州無法享受的幸福。</p><p>雖然沒有住在灣區的摩登公寓，但在這個城市，我住得相對有歸屬感。</p><p>來到上海工作一個多月後回過頭看，我沒有一絲後悔，甚至有些慶幸：<strong>自己忍痛主動放棄抱了三年的夢想、拋下了在美國四年的積累，收穫是 — — 換了條路，離我真正想到達的地方，卻更近了些。</strong></p><p>Feel free to leave a comment below or DM me on Instagram! <a href="https://www.instagram.com/chenjing.t.w/">@chenjing.t.w</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7d7044bcd94" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A project I did as a data science intern: Solar-Powered House Data Analysis Report]]></title>
            <link>https://medium.com/@chenjing.claire/https-medium-com-chingchen-17-data-analysis-report-solar-powered-house-7c915ac10132?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/7c915ac10132</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[solar-energy]]></category>
            <category><![CDATA[internships]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Thu, 27 Jul 2017 10:27:39 GMT</pubDate>
            <atom:updated>2017-07-27T10:41:14.596Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zVBc7lf_SmllMg2fMWLL7A.png" /></figure><p><em>*This project was done during my internship at </em><a href="https://mobagel.com/tw/"><em>Mobagel</em></a><em>.</em></p><p>There is a solar-powered house located in Taiwan. With the given dataset, my goal is to predict and propose the following:</p><ol><li>Predict the daily pattern of household electricity usage of each month</li><li>Predict the daily pattern of solar-power generation of each month</li><li>Ideal inverter power</li><li>Ideal peak household electricity consumption</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*cU-k1u5xOgyoZ-0SeLZh_g.png" /><figcaption>Credit: AGL Solar</figcaption></figure><h3>How solar-powered house works</h3><p>On sunny days, solar panels convert sunlight to DC electricity and part of it is stored in the battery. The inverter converts DC electricity to AC electricity to supply the house loads. Appliances in the house will draw electricity from the inverter first. When there is surplus, it gets exported to the grid (mains). But when there isn’t enough electricity to supply the house loads, it will top it up by importing more electricity from the grid (mains).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/650/1*glipvh44OGtCHIjvhPXKQQ.png" /></figure><p><strong>Data Visualization</strong>: ‘house_KW’</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IFEtv5-paxBHlYY_bSsD0g.png" /></figure><ul><li>Found out that the November dataset is abnormal,</li><li>and the power consumption is higher in summer than in winter</li></ul><p><strong>Daily pattern: </strong>(x = _id.hour, y = _id.month)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*bQmBt0IRNO1Iu9AMijCz1A.png" /></figure><p>Battery percentage peak: <br>10:00 ~ 17:00</p><p>House_KW peak: 20:00 ~ midnight</p><p>House_KWH: daily cumulative usage</p><p>Mains — sell: 10:00 ~ 16:00; buy: 16:00 ~10:00</p><p>Solar_powergen: high in August, low in November</p><p>Solar_stats: accumulative power generation</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*YkXX3-F_A2jxzsaov5WZzQ.png" /></figure><p>No matter how much power the battery has, it often imports electricity from the grid, and rarely exports.</p><ol><li><strong>Predict the daily pattern of solar-power generation of each month</strong></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/362/1*PwIDyVNFm6bSa8BNdIxZJg.png" /></figure><p>September’s hourly solar_powergen <br>(X = time, y = solar_powergen)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*JheftyFAdwalGHKRHdjZuQ.png" /></figure><p>September’s average daily solar_powergen</p><p><strong>Feature Engineering</strong>: <br>1. Add features including sunshine hours, cloudiness, temperature, number of days temperature higher than 30 degrees Celsius. I chose these features because solar-power generation must be highly affected by those factors. (Data source: <a href="http://www.cwb.gov.tw/V7/climate/dailyPrecipitation/dP.htm">交通部中央氣象局</a>）</p><p>2. I grouped data by month and hour, since I presumed knowing daily pattern of each month is sufficient in this case. And I got the median values of each columns to make sure that they are not influenced by outliers.</p><p><strong>Model:</strong></p><p>I have tried Random forest, gradient boosting and bagging regressor (because they often have better performance than other estimators), and it turned out that gradient boosting regressor, on the test set, has the highest R² score among all. <br><em>*(Bagging: 0.793/ random forest: 0.807/ gradient boosting: 0.816)</em></p><p>GradientBoostingRegressor(max_depth=3, learning_rate=0.1, n_estimators=200, subsample=0.9, random_state=0)</p><p>X_train = Time columns+ hourly average of the rest of the columns <br>y_train = solar_powergen<br>X_test = Time columns + hourly average of the rest of the columns from the training set <br>y_test = solar_powergen</p><p>*Note: “Accuracy” in this report is defined as (1- | sum of the absolute value of prediction- true / sum of prediction|)%</p><ul><li>The gbr.score (R² score) below is the score of each sample and its prediction.</li><li>The accuracy below represents how close the predictions and the true values of average daily solar-power generation of each month is.</li></ul><blockquote>R² provides a measure of how well observed outcomes are replicated by the model, based on the proportion of total variation of outcomes explained by the model . Accuracy is a description of observational errors. (<a href="https://en.wikipedia.org/wiki/Coefficient_of_determination">Wiki</a>)</blockquote><p>R² measures “explained variance” but not prediction error, so I used two evaluations to show how well the model performs and how accurate the prediction is.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*sEwtCoM5NdP5IADM_WJKwQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*gek9bBjXxLHDKJUUCLqr_w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*kxiNCXwJbDzEZ1OFjpaK4g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*PTVWDwL2k2X3EvrVVGl7qA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*VgO2b7BEpx0gONurW0_WkQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*U7BaL1SDj70cXPE_ie8nPg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*1JvUfxAdlDBPZHj5rUaPTg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*p5xt1gcIt2aoQi7PXGf5oA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*wXoQWrisdfZdaOTFaeQm1g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*9d8l_rWNsDfpJ7Es9GqmqA.png" /></figure><pre>month: 8<br>gbr.score:  0.949485042148<br>Accuracy:  0.87971514877</pre><pre><br>month: 9<br>gbr.score:  0.816181749495<br>Accuracy:  0.783695311456</pre><pre>month: 10<br>gbr.score:  0.791087363623<br>Accuracy:  0.667679129653</pre><pre>month: 11<br>gbr.score:  0.884837348683<br>Accuracy:  0.71054816527</pre><pre>month: 12<br>gbr.score:  0.927028536038<br>Accuracy:  0.895645958349</pre><pre>month: 1<br>gbr.score:  0.833262924263<br>Accuracy:  0.767611800185</pre><pre><br>month: 2<br>gbr.score:  0.896935738999<br>Accuracy:  0.817152968986</pre><pre>month: 3<br>gbr.score:  0.758768541144<br>Accuracy:  0.779393732883</pre><pre>month: 4<br>gbr.score:  0.856984796955<br>Accuracy:  0.722903100651</pre><pre>month: 5<br>gbr.score:  0.859756340341<br>Accuracy:  0.820325939924</pre><p>Average accuracy of each month grouped by hour = 0.784<br>Average R² score of each month grouped by hour = 0.857</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/1*ZtLc_SINH_1am2VDU9-Ofw.png" /><figcaption>cross_val_predict (cv=10)</figcaption></figure><p><strong>2. Predict the daily pattern of household electricity usage of each month</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*jSs6yiJv0R_DUxoI52xGpg.png" /></figure><p>September’s hourly household power consumption<br>(X = time, y = house_KW)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*SE9XIh1CaOukNU7E7CJXdw.png" /></figure><p>September’s daily household power consumption</p><p><strong>Model: <br></strong>GradientBoostingRegressor(<br> max_depth=2,<br> learning_rate=0.09, <br> n_estimators=50, <br> random_state=0)</p><p>X_train = Time columns+ hourly average of the rest of the columns <br>y_train = house_KW<br>X_test = Time columns + hourly average of the rest of the columns from the training set <br>y_test = house_KW</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/377/1*9q1_mzsi5p0Sv6FZmToxSA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*rzXipTYcSlmtIna1nplxtw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*8JgIpdC5J64D2Fn3ufNyuw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*ZQAbJr7XBaho0akkaUdl8g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*4HbNUgm8UI88eXF0oV-AAA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*GTm9QtaSaaO_BpU4sj3fpw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*uvayBxgEgcL5kaCqdtLSVA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*SgO6Gom0n9tKjdxO_GKZqQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*Dann0gAzHNYxuE-BpMmnFg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/372/1*Zmz3PRioFzQDHQI0V-meDA.png" /></figure><pre>month: 8<br>Accuracy of each sample: 0.730<br>Accuracy of daily average:0.754</pre><pre>month: 9<br>Accuracy of each sample: 0.534<br>Accuracy of daily average:0.617</pre><pre>month: 10<br>Accuracy of each sample: 0.704<br>Accuracy of daily average:0.798</pre><pre>month: 11<br>Accuracy of each sample: 0.650<br>Accuracy of daily average:0.711</pre><pre>month: 12<br>Accuracy of each sample: 0.730<br>Accuracy of daily average:0.791</pre><pre>month: 1<br>Accuracy of each sample: 0.667<br>Accuracy of daily average:0.740</pre><pre>month: 2<br>Accuracy of each sample: 0.778<br>Accuracy of daily average:0.813</pre><pre>month: 3<br>Accuracy of each sample: 0.731<br>Accuracy of daily average:0.777</pre><pre>month: 4<br>Accuracy of each sample: 0.651<br>Accuracy of daily average:0.751</pre><pre><br>month: 5<br>Accuracy of each sample: 0.547<br>Accuracy of daily average:0.620</pre><p>On test set: <br>Average accuracy of each month = 0.672<br>Average accuracy of each month grouped by hour = 0.737<br>Average CV Score of each month’s median CV score (exclude November)= 0.639</p><p>On full dataset:<br>Average accuracy of each month = 0.68<br>Average CV score of each month’s median CV score (exclude November)= 0.627</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/386/1*hcccIEoo6K3mNm2futV5kw.png" /><figcaption>cross_val_predict (cv=10)</figcaption></figure><p>Benchmark:<br>Based on pure statistics (calculated by the median values of house_KW of each hour in each month), the accuracy of each sample is 0.48, and that of the daily average value is 0.62. <br>The performance of my model is better than the benchmark.</p><p><strong>3. Ideal inverter power<br></strong>Model: <br>DecisionTreeClassifier(max_depth=3, random_state=1, <br>min_samples_split=0.1, max_leaf_nodes=5, min_samples_leaf= 0.1)<br>cv_score = <strong>0.874</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/459/1*d5eRy2Ap0W68KQN-1KSlEw.png" /></figure><p>The house can mostly operate independently without importing electricity from the grid when:</p><ol><li>the inverter power is larger than 0.838KW (which accounts for about 20% of time)</li><li>the inverter is smaller than 0.838KW and house_KW smaller than about 0.23KW (which accounts for about 10% of time)</li></ol><p>Current Inverter Power：<br>mean 0.533212 ｜ std 0.796064<br>min -1.945600 ｜max 4.410000<br>25% 0.000000｜50% 0.237000｜75% 0.711000</p><p>The total amount of electricity imported from the mains is 1875.57kWh, which costs NTD 3923.07. <br>The total amount of electricity exported to the mains is 1740.60kWh, and made NTD 12,463 (calculated in 7.16NTD/kWh ).（Data Source：<a href="http://www.taipower.com.tw/UpFile/_userfiles/file/%E8%A1%A8%E7%87%88%E9%9D%9E%E6%99%82%E9%96%93%E9%9B%BB%E5%83%B9%E8%A8%88%E7%AE%97%E7%AF%84%E4%BE%8B_105.04.01.pdf">台電電價計算範例</a>）</p><p>Under the condition that battery has charge left, if we can improve the inverter power to the 75th percentile of house_KW (0.7KW), we can reduce the total imported amount of electricity to 185.79 KW, and decrease the fee to NTD 302.83. Moreover, we can sell 1055.7kWh more back to the grid, and make NTD 20,021.</p><p>The most ideal inverter power is 0.838 KW. With that, the solar-powered house would be very close to a energy self-sufficient one.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/329/1*novIFYg4C_fuqr9YQtaCEw.png" /></figure><p>The correlation b/w inverter and solar_powergen is high.</p><p>Linear Regression:</p><pre>Coefficients: [[ 0.661]]<br>Intercept: [ 0.128]<br>Score: 0.712</pre><p>inverter = 0.128 + 0.661*solar_powergen</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F_uLGYih7Y1II1CMmEkapA.png" /></figure><p>Suggestion: Better use appliances during daytime (10:00~16:00) when the inverter has more power to support so that we can reduce the amount of power imported from the grid.</p><p><strong>4. Ideal peak household electricity consumption<br></strong>Model:<br>DecisionTreeClassifier(max_depth=3, random_state=1, min_samples_split=0.001, max_leaf_nodes=5, min_samples_leaf= 0.1)<br>cv_score = <strong>0.828</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/475/1*MldGI1H25X8P98vxXcKKCQ.png" /></figure><p>Most of the time when solar_powergen is larger than 1.467KW (which accounts for about 20% of time), there is no restriction on household electricity consumption. Otherwise, better limit the household electricity consumption below 0.23kW to keep independent from the mains.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7c915ac10132" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[When AI meets 3000-year-old Chinese Palmistry]]></title>
            <link>https://medium.com/data-science/when-ai-meets-3000-year-old-chinese-palmistry-a767b7f3defb?source=rss-c0703fd48882------2</link>
            <guid isPermaLink="false">https://medium.com/p/a767b7f3defb</guid>
            <category><![CDATA[hackathons]]></category>
            <category><![CDATA[palm-reading]]></category>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Chenjing.Claire]]></dc:creator>
            <pubDate>Wed, 26 Jul 2017 09:34:58 GMT</pubDate>
            <atom:updated>2017-08-03T21:27:52.972Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/692/1*aQKlMG9k8BhOinKiNbuNzQ.png" /></figure><h4><a href="https://www.facebook.com/handbot2017">Handbot</a> — the first palm reading chatbot using deep learning (CNN) to analyze one’s palm to predict his/her character traits, health, career, relationship and many more aspects.</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Hadqeb8eH6SZPLq-5QS4ng.jpeg" /></figure><p>In a 48-hour hackthon (HackNTU 2017 @ Taipei, Taiwan), two of my coworkers from the startup I’m interning at, a college friend of mine, my brother, and I cooperatively built this first palm reading chatbot, aka <a href="https://www.facebook.com/handbot2017">Handbot</a>. We aspire to bring a disruptive innovation into the 3000-year-old Chinese Palmistry industry.</p><p>Fortunately, we won an information security award given by Hitcon.</p><p>If you are a devotee of science and struggle to believe fortune telling, here is a more scientific explanation that might change your mind —</p><blockquote>“The human hand contains a wealth of information, because a baby’s hands form early in gestation, researchers like to say that they amount to a “fossilized record” of early development, one that may provide insight into future well-being.”<br> — “Palm Reading Is Real?”, <a href="https://www.theatlantic.com/magazine/archive/2015/07/palm-reading/395288/">The Atlantic.</a></blockquote><p>As we know, a lot of people are obsessed with fortune telling, and palm reading is one of the easiest methods that everyone is able to get started in foreseeing their fortune. However, many people still have problems with identifying the lines of their palm. Therefore, we came up with this idea: combining “<strong>palm reading</strong>”, “<strong>deep learning</strong>”, and “<strong>chatbot.</strong>”</p><p>Palm reading: We spent the first couple of hours at the hackathon to learned basic palmistry including relationship, career and health analyzation and prediction. Then, we labeled over 2000 pictures of real human palms and gave over 6000 scores of their relationship, career and health respectively. (You can find our label file <a href="https://github.com/kkshyu/palm-read/blob/master/labels.csv">here</a>.)</p><p>Deep learning: We use VGG-16(16-layer network) in Keras with Tensorflow backend as our CNN model. Using GPU to accelerate, we have trained 10 epoches and 1000 steps per epoch. The MSE of the model is 1.3066. Besides, the valuation MSE is 1.1721. Overall, the result is robust enough to predict a new palm. (You can find our code <a href="https://github.com/kkshyu/palm-read/blob/master/train_palm.py">here</a>.)</p><p>Chatbot, compared with App and Web, is more accessible to mobile users, and it is also easier to spread on social media. Therefore, we used <a href="https://chatfuel.com/">Chatfuel</a> to create an AI chatbot on Facebook Messenger. Below you can see the user flow of Handbot. The smooth UX we designed help us achieve <strong>80%</strong> user retention rate.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hYzJLZMJ8I2yFWieF-XjbA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1013/1*iEhC8KiTYV74xcbLQ0P-PA.png" /></figure><p><em>Over 80 users within a day only with personal Facebook sharing.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1020/1*DAlbtvDVk-lFdQL8SNTA7A.png" /></figure><p><em>Over 70 active users, and most of the users had pressed a button after they got started in Handbot.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Wx9INipGpaGm6NsbDnZdEg.png" /></figure><p><em>You can see our Demo on HITCON’s streaming video </em><a href="https://www.facebook.com/HITCON/videos/1558077734223543/"><em>here</em></a><em> (starting from 1:48:00). <br>You can also read the news report about the winners of this hackathon (including us) </em><a href="http://campus.chinatimes.com/20170724004521-262301"><em>here</em></a><em>.</em></p><h4>Hidden security problems of chatbot applications</h4><p>The first night we launched Handbot at the hackathon, and promoted on social media, we not only received messages from our friends, but also some HackNTU attendees and some strangers. We started getting pictures of their palms (sometime their faces), the backgrounds behind the palms, and moreover, their personal information on Facebook. We were shocked by how much users’ information we could obtain from this chatbot. We could know who they are, where they were, and even their fingerprints!</p><p>People are less aware of information security risks when using application like chatbot: we are so used to casually chat with friends on Messenger, so we got a mindset that <strong>“conversations happened on Messenger is safe and private.”</strong> We tend to treat a chatbot as a friend and we have no filter texting it.</p><p>Global Chatbot market was expected to grow exponentially during the forecast period 2016–2023. As users, we should be more aware of the hidden security problems. Giant social media platforms like Facebook should actively intercept inappropriate applications.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a767b7f3defb" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/when-ai-meets-3000-year-old-chinese-palmistry-a767b7f3defb">When AI meets 3000-year-old Chinese Palmistry</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>