<?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 Thomas Dittmar on Medium]]></title>
        <description><![CDATA[Stories by Thomas Dittmar on Medium]]></description>
        <link>https://medium.com/@bookmemate?source=rss-7d69d58a320f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ASTAzd6hD7c-rC1x4sYi7w.png</url>
            <title>Stories by Thomas Dittmar on Medium</title>
            <link>https://medium.com/@bookmemate?source=rss-7d69d58a320f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 17:13:53 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@bookmemate/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[Deep Learning with React Native (iOS only)]]></title>
            <link>https://medium.com/data-science/deep-learning-with-react-native-ios-only-8089fed59773?source=rss-7d69d58a320f------2</link>
            <guid isPermaLink="false">https://medium.com/p/8089fed59773</guid>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[tensorflow]]></category>
            <category><![CDATA[react-native]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Thomas Dittmar]]></dc:creator>
            <pubDate>Thu, 13 Aug 2020 11:18:38 GMT</pubDate>
            <atom:updated>2020-08-17T03:52:12.077Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dm0rZczn-IfHwL1o7rabUw.jpeg" /><figcaption>(Image by Author)</figcaption></figure><p><em>Article originally published on </em><a href="https://dev.to/dittmarconsulting/deep-learning-with-react-native-ios-only-2470"><em>dev.to</em></a></p><h4>Intro</h4><p>In this tutorial, I will cover all the steps on how to build a mobile application and train a deep learning model so that you can predict a handwritten digit between 0 and 9 by using your mobile’s camera.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/443/1*Bk4GS5ff5r2uSUfknBfZDw.gif" /><figcaption>App predicting the handwritten number (Image by Author)</figcaption></figure><p>But before we start building the mobile app, we need to come up with a high-level strategy. Let’s go through the thought process:</p><ul><li>Do we build a bare React-Native (<strong>RN</strong>) or an <strong>Expo</strong> app?</li><li>Which camera library do we want to use?</li><li>Do we have to crop the images and what library we want to use?</li><li>How do we train a deep learning model?</li><li>How do we use that model against the photo?</li><li>How do we show the result?</li></ul><blockquote>Note: This tutorial needs some prerequisites and a good understanding of RN and Javascript in general. If you are an absolute beginner, I would suggest following a good course on Youtube, Udemy or Egghead before continuing with this tutorial.</blockquote><h4>Let’s get started</h4><p>I’m going to divide this tutorial into three sections</p><p><a href="#1a09"><strong>Chapter 1</strong>: Create the RN application</a><br><a href="#c7cf"><strong>Chapter 2</strong>: Train the deep learning model</a><br><a href="#01f8"><strong>Chapter 3</strong>: Implement the model, predict and show the result</a></p><h4>Chapter 1: Create the RN application</h4><p>Remember the first point of our thought process whether to create a bare or an Expo boilerplate app?</p><p>After some research, I decided to load the trained model locally. It’s the easiest way without fetching the model from a cloud server, but you can do that too.</p><p>In this tutorial, we will use bundleResourceIO from <a href="http://twitter.com/tensorflow/tfjs-react-native">@tensorflow/tfjs-react-native</a> that is unfortunately not compatible with Expo.</p><blockquote>Also, since we want to use the camera we have to use a physical phone and not a simulator. For that you must have an Apple Developer Account to sign your app otherwise you will not be able to run the app.</blockquote><p>Let’s create the app with the following command:</p><pre>$ react-native init MyFirstMLApp</pre><p>After the installation process has been completed make sure all your pods are installed too!</p><pre>$ cd MyFirstMLApp<br>$ npx pod-install</pre><p>Let’s run the app for the first time on your physical iPhone. Open Xcode and find the MyFirstMLApp.xcworkspace and open it. Connect your iPhone to your Mac with the lightning cable and select your phone. Press the play button for building and running the app for the first time. You should see the <strong>Welcome to React </strong>screen on your iPhone.</p><p>🏆 <strong>Awesome!</strong></p><p>Let’s add some packages for this app:</p><pre>yarn add <a href="http://twitter.com/react">@react</a>-native-community/async-storage <a href="http://twitter.com/react">@react</a>-native-community/cameraroll <a href="http://twitter.com/tensorflow/tfjs">@tensorflow/tfjs</a> <a href="http://twitter.com/tensorflow/tfjs-react-native">@tensorflow/tfjs-react-native</a> expo-camera expo-gl expo-gl-cpp expo-image-manipulator react-native-fs react-native-svg react-native-unimodules victory-native</pre><p>and finally, install the navigation library.</p><pre>yarn add react-native-navigation &amp;&amp; npx rnn-link</pre><p>The latter command will add the navigation package to iOS and Android. But we are not quite just done yet.</p><p>Because we use the bare framework of RN, <a href="https://docs.expo.io/bare/installing-unimodules/">unimodules</a> need to be installed manually.</p><p>Please click on the link and modify the Podfile as described in the iOS section. After that run</p><pre>$ npx pod-install</pre><p>and build the Xcode project to see if everything has been installed correctly.</p><p>Then continue with adding the code for the <a href="https://docs.expo.io/bare/installing-unimodules/">unimodules</a> to the AppDelegate.m and build the project again.</p><p>Because we want to use the camera to take pictures we also need to add a few privacy keys to the Info.plist</p><pre>&lt;?xml version=”1.0&quot; encoding=”UTF-8&quot;?&gt;<br>&lt;!DOCTYPE plist PUBLIC “-//Apple//DTD PLIST 1.0//EN” “<a href="http://www.apple.com/DTDs/PropertyList-1.0.dtd">http://www.apple.com/DTDs/PropertyList-1.0.dtd</a>&quot;&gt;<br>&lt;plist version=”1.0&quot;&gt;<br>  &lt;dict&gt;<br>    <br>    &lt;! — Required for iOS 10 and higher --&gt;<br>    &lt;key&gt;NSCameraUsageDescription&lt;/key&gt;<br>    &lt;string&gt;We need to use the camera for taking pictures of the digits&lt;/string&gt;</pre><pre>&lt;! — Required for iOS 11 and higher: include this only if you are planning to use the camera roll --&gt;<br>    &lt;key&gt;NSPhotoLibraryAddUsageDescription&lt;/key&gt;<br>    &lt;string&gt;We need to access the photo library to upload the images&lt;/string&gt;</pre><pre>&lt;! — Include for only if you are planning to use the camera roll --&gt;<br>    &lt;key&gt;NSPhotoLibraryUsageDescription&lt;/key&gt;<br>    &lt;string&gt;We need to access the photo library to upload the images&lt;/string&gt;</pre><pre>&lt;! — Include this only if you are planning to use the microphone for video recording --&gt;<br>    &lt;key&gt;NSMicrophoneUsageDescription&lt;/key&gt;<br>    &lt;string&gt;We need to access the microphone&lt;/string&gt;<br> <br>    &lt;key&gt;CFBundleDevelopmentRegion&lt;/key&gt;<br>    &lt;string&gt;en&lt;/string&gt;</pre><p>If Xcode builds fine you can either continue running the app from Xcode or just use the terminal.</p><p>If you decide to run the app from the command line from now on, like me, please add — device to the ios script in your package.json file and run</p><pre>yarn ios</pre><p>Once the app is starting on your iPhone don’t be surprised that you don’t see the welcome page anymore. That’s because we use react-native-navigation. But you should see the loading screen <strong>MyFirstMLApp</strong></p><p>Now it’s time to create our 2 only screens and add the navigation for these screens to our project.</p><p>Please create the src/screens/CameraView and src/screens/EvaluationView directories in the root of our project.</p><p>Inside src/screens/CameraView create an index.js file and add this code:</p><pre>import React, { useState, useRef, useEffect } from &quot;react&quot;;<br>import {<br>  SafeAreaView,<br>  TouchableOpacity,<br>  View,<br>  Text,<br>  StatusBar,<br>} from &quot;react-native&quot;;<br>import { Navigation } from &quot;react-native-navigation&quot;;<br>import { Camera } from &quot;expo-camera&quot;;</pre><pre>const MASK_DIMENSION = 100;</pre><pre>export const CameraView = (props) =&gt; {<br>  const [hasPermission, setHasPermission] = useState(null);<br>  const [showShutterButton, setShowShutterButton] = useState(false);<br>  const cameraRef = useRef();</pre><pre>useEffect(() =&gt; {<br>    (async () =&gt; {<br>      const { status } = await Camera.requestPermissionsAsync();<br>      setHasPermission(status === &quot;granted&quot;);<br>    })();<br>  }, []);</pre><pre>const handlePictureProcessing = async () =&gt; {<br>    goToEvaluationView();<br>  };</pre><pre>const goToEvaluationView = () =&gt; {<br>    Navigation.push(props.componentId, {<br>      component: {<br>        name: &quot;evaluationView&quot;,<br>        options: {<br>          topBar: {<br>            title: {<br>              text: &quot;Evaluating ML result&quot;,<br>              color: &quot;white&quot;,<br>            },<br>            background: {<br>              color: &quot;#4d089a&quot;,<br>            },<br>            backButton: {<br>              color: &quot;white&quot;,<br>              showTitle: false,<br>            },<br>          },<br>        },<br>        passProps: {},<br>      },<br>    });<br>  };</pre><pre>if (hasPermission === null) {<br>    return &lt;View /&gt;;<br>  }</pre><pre>if (hasPermission === false) {<br>    return &lt;Text&gt; No access to camera &lt;/Text&gt;;<br>  }</pre><pre>return (<br>    &lt;React.Fragment&gt;<br>      &lt;StatusBar barStyle=&quot;light-content&quot; /&gt;<br>      &lt;SafeAreaView style={styles.safeArea}&gt;<br>        &lt;Camera<br>          ref={cameraRef}<br>          type={Camera.Constants.Type.back}<br>          whiteBalance={Camera.Constants.WhiteBalance.auto}<br>          onCameraReady={() =&gt; setShowShutterButton(true)}&gt;<br>          &lt;View style={styles.cameraView}&gt;<br>            &lt;View style={styles.mask} /&gt;<br>            {showShutterButton &amp;&amp; (<br>              &lt;TouchableOpacity<br>                style={styles.shutterButton}<br>                onPress={handlePictureProcessing}&gt;<br>                &lt;Text style={styles.shutterButtonText}&gt;<br>                  Take a picture<br>                &lt;/Text&gt;<br>              &lt;/TouchableOpacity&gt;<br>            )}<br>          &lt;/View&gt;<br>        &lt;/Camera&gt;<br>      &lt;/SafeAreaView&gt;<br>    &lt;/React.Fragment&gt;<br>  );<br>};</pre><pre>const styles = {<br>  safeArea: {<br>    backgroundColor: &quot;#4d089a&quot;,<br>  },<br>  cameraView: {<br>    height: &quot;100%&quot;,<br>    justifyContent: &quot;center&quot;,<br>    alignItems: &quot;center&quot;,<br>    backgroundColor: &quot;transparent&quot;,<br>  },<br>  mask: {<br>    height: MASK_DIMENSION,<br>    width: MASK_DIMENSION,<br>    borderWidth: 3,<br>    borderColor: &quot;white&quot;,<br>    borderStyle: &quot;dotted&quot;,<br>    borderRadius: 15,<br>  },<br>  shutterButton: {<br>    position: &quot;absolute&quot;,<br>    bottom: 0,<br>    width: 150,<br>    height: 40,<br>    justifyContent: &quot;center&quot;,<br>    alignItems: &quot;center&quot;,<br>    borderWidth: 1,<br>    borderColor: &quot;white&quot;,<br>    borderRadius: 15,<br>    marginBottom: 20,<br>  },<br>  shutterButtonText: {<br>    fontSize: 18,<br>    color: &quot;white&quot;,<br>  },<br>};</pre><pre>CameraView.options = {<br>  statusBar: {<br>    backgroundColor: null,<br>  },<br>  topBar: {<br>    title: {<br>      text: &quot;Take a picture&quot;,<br>      color: &quot;white&quot;,<br>    },<br>    background: {<br>      color: &quot;#4d089a&quot;,<br>    },<br>  },<br>  tapBar: {<br>    background: {<br>      color: &quot;#4d089a&quot;,<br>    },<br>  },<br>};</pre><p>Inside src/screens/EvaluationView create an index.js file and add this code:</p><pre>import React from &quot;react&quot;;<br>import { SafeAreaView, View, Text, StatusBar } from &quot;react-native&quot;;</pre><pre>export const EvaluationView = (props) =&gt; {<br>  return (<br>    &lt;React.Fragment&gt;<br>       &lt;StatusBar barStyle=&quot;light-content&quot; /&gt;<br>      &lt;SafeAreaView style={styles.safeArea}&gt;<br>        &lt;View style={styles.container}&gt;<br>           &lt;Text style={styles.headerText}&gt;ANALYSIS&lt;/Text&gt;<br>        &lt;/View&gt;<br>      &lt;/SafeAreaView&gt;<br>    &lt;/React.Fragment&gt;<br>  );<br>};</pre><pre>const styles = {<br>  safeArea: {<br>    backgroundColor: &quot;#4d089a&quot;,<br>  },<br>  container: {<br>    height: &quot;100%&quot;,<br>    alignItems: &quot;center&quot;,<br>    backgroundColor: &quot;white&quot;,<br>  },<br>  headerText: {<br>    fontSize: 20,<br>    fontWeight: &quot;500&quot;,<br>    color: &quot;#4d089a&quot;,<br>    margin: 20,<br>  },<br>};</pre><p>Then override the index.js file in your root with the following code:</p><pre>import { Navigation } from &quot;react-native-navigation&quot;;<br>import { CameraView } from &quot;./src/screens/CameraView&quot;;<br>import { EvaluationView } from &quot;./src/screens/EvaluationView&quot;;</pre><pre>Navigation.registerComponent(&quot;cameraView&quot;, () =&gt; CameraView);<br>Navigation.registerComponent(&quot;evaluationView&quot;, () =&gt; EvaluationView);</pre><pre>Navigation.setDefaultOptions({<br>  statusBar: {<br>    style: &quot;light&quot;,<br>    backgroundColor: &quot;#4d089a&quot;,<br>  },<br>  topBar: {<br>    title: {<br>      color: &quot;white&quot;,<br>    },<br>    background: {<br>      color: &quot;#4d089a&quot;,<br>    },<br>    backButton: {<br>      color: &quot;white&quot;,<br>      showTitle: false,<br>    },<br>  },<br>});</pre><pre>Navigation.events().registerAppLaunchedListener(() =&gt; {<br>  Navigation.setRoot({<br>    root: {<br>      stack: {<br>        children: [<br>          {<br>            component: {<br>              name: &quot;cameraView&quot;,<br>            },<br>          },<br>        ],<br>      },<br>    },<br>  });<br>});</pre><p>Finally, you can remove the App.js file as it’s not needed anymore.</p><p>Restart your metro bundler and you should see the app running like that …</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/443/1*MnLgHukfbKevE2GqvUJ5fw.gif" /><figcaption>Screen recording of the app with working navigation (Image by Author)</figcaption></figure><p>🏆 <strong>Congratulations!</strong></p><p>You have created the base app that doesn’t take pictures yet but can navigate from one screen to the other.</p><h4>Chapter 2: Train the deep learning model</h4><p>Initially, I used <a href="https://www.kaggle.com/c/digit-recognizer/data">this pre-trained model from <strong>Kaggle</strong></a><strong> </strong>but the effort to make the app work was huge.</p><p>I had to create created an <strong>AWS EC2 Deep Learning AMI (Amazon Linux 2) Version 30.1 instance</strong> with SSH access because my Macbook doesn’t support CUDA. (GPU support needed to train the model)<br>Then I had to copy the <a href="https://jupyter.org/"><strong>Jupyter</strong></a><strong> </strong>notebook from Kaggle, ran the notebook to train the model on the AWS instance (it ran for 3 hours) and moved the model back into my project.<br>Furthermore, I had to install <strong>OpenGL</strong> to modify the image and wrote a quite complex script to reshape the base64 string to a tensor to match the expected input for the model [1, 28, 28, 1].</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/606/1*tEa0cwfoHxvWm6Sm6jgOFw.gif" /><figcaption>The model started training on AWS (Image by Author)</figcaption></figure><p>All that made me rethink how to write this tutorial. After all, this tutorial should be for people that just want to play around with the machine learning model without learning <strong>Python</strong>, <strong>Jupyter</strong>, <strong>Tensorflow</strong> and <strong>Keras</strong> beforehand. Also, the tutorial would have been 5 times the length of what it is now.</p><blockquote>Note: If you want to learn how to use <strong>Tensorflow</strong> &amp; <strong>Keras</strong> I found a good Youtube channel about deep learning with <a href="https://www.youtube.com/channel/UC4UJ26WkceqONNF5S26OiVw"><strong>deeplizard</strong> </a>, which is very informative and it is in line with what we want to do in this tutorial.<br>Also, <a href="https://www.udemy.com/course/complete-machine-learning-and-data-science-zero-to-mastery/learn/lecture/16488468"><strong>this course on Udemy</strong></a><strong> </strong>is pretty good but it’s not free unfortunalety. <em>😕</em></blockquote><p>Anyway, for this tutorial, I decided to use <a href="https://teachablemachine.withgoogle.com/train/image"><strong>Google’s Teachable Machine</strong></a> to train the images.</p><p>The idea is to take <strong>28 x 28 pixels</strong> images with the app we just build, upload the images to the Teachable Machine and download the trained model back into our project.</p><blockquote>Just in case you were asking why I use 28 x 28 pixels images? Well, that was the original input size of the model I used first. So I stuck with it.</blockquote><p>That also means we have to crop and save the taken images to the camera library. In order to do that we need to modify our code a little.</p><p>Please create a helper.js file inside the CameraView folder and paste the following code:</p><pre>import { Dimensions } from &quot;react-native&quot;;<br>import * as ImageManipulator from &quot;expo-image-manipulator&quot;;<br>import CameraRoll from &quot;<a href="http://twitter.com/react">@react</a>-native-community/cameraroll&quot;;</pre><pre>const { height: DEVICE_HEIGHT, width: DEVICE_WIDTH } = Dimensions.get(&quot;window&quot;);</pre><pre>// got the dimension from the trained data of the *Teachable Machine*; pixel resolution conversion (8x)<br>export const BITMAP_DIMENSION = 224;</pre><pre>export const cropPicture = async (imageData, maskDimension) =&gt; {<br>  try {<br>    const { uri, width, height } = imageData;<br>    const cropWidth = maskDimension * (width / DEVICE_WIDTH);<br>    const cropHeight = maskDimension * (height / DEVICE_HEIGHT);<br>    const actions = [<br>      {<br>        crop: {<br>          originX: width / 2 - cropWidth / 2,<br>          originY: height / 2 - cropHeight / 2,<br>          width: cropWidth,<br>          height: cropHeight,<br>        },<br>      },<br>      {<br>        resize: {<br>          width: BITMAP_DIMENSION,<br>          height: BITMAP_DIMENSION,<br>        },<br>      },<br>    ];<br>    const saveOptions = {<br>      compress: 1,<br>      format: ImageManipulator.SaveFormat.JPEG,<br>      base64: false,<br>    };<br>    return await ImageManipulator.manipulateAsync(uri, actions, saveOptions);<br>  } catch (error) {<br>    console.log(&quot;Could not crop &amp; resize photo&quot;, error);<br>  }<br>};</pre><pre>export const saveToCameraRoll = async (uri) =&gt; {<br>  try {<br>    return await CameraRoll.save(uri, &quot;photo&quot;);<br>  } catch (error) {<br>    console.log(&quot;Could not save the image&quot;, error);<br>  }<br>};</pre><p>Add the import this file in src/screens/CameraView/index.js</p><pre>import { cropPicture, saveToCameraRoll } from ‘./helpers’;</pre><p>and add the takePicture function and modify the handlePictureProcessing function like that</p><pre>const handlePictureProcessing = async () =&gt; {<br>  const imageData = await takePicture();<br>  const croppedData = await cropPicture(imageData, MASK_DIMENSION);<br>  await saveToCameraRoll(croppedData.uri);<br>  // we don&#39;t want to go to the evaluation view now<br>  //goToEvaluationView();<br>};</pre><pre>const takePicture = async () =&gt; {<br>  const options = {<br>    quality: 0.1,<br>    fixOrientation: true,<br>  };<br>  try {<br>    return await cameraRef.current.takePictureAsync(options);<br>  } catch (error) {<br>    console.log(&quot;Could not take photo&quot;, error);<br>  }<br>};</pre><p>As you can see we comment out the line //goToEvaluationView(); so we don’t go to the other screen. That means you can take as many pictures as you want in a row. All the pictures will now be saved in your photo library.</p><p>Our next task is to write as many variations of numbers between 0 and 9 as possible on a piece of paper. The more numbers, colours and pen shapes we use the better the prediction will be.</p><p>I was lazy and ended up with roughly 10 variations per number but the prediction was a bit off for a few numbers such as 4 and 8.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cVMikcy1zDrnxif2UV9dmw.jpeg" /><figcaption>Wrote variations of number shapes (Image by Author)</figcaption></figure><p>So it’s up to you how many numbers you will let the <strong>Teachable Machine</strong> train on.</p><p>When you finished with taken the images, Airdrop them all back to your Mac and from there upload them to the <a href="https://teachablemachine.withgoogle.com/train/image"><strong>Teachable Machine</strong></a> and start training them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oPs5sYmr8xNV73f8ncbEMg.png" /><figcaption>Screenshot of the trained model (Image by Author)</figcaption></figure><p>Once that has finished you can take additional pictures with your app and upload them too, to test against the trained model.</p><p>If you are happy with the result, click on Export Model -&gt; Tensorflow.js -&gt; Download -&gt; Download my model, which will download a <strong>ZIP</strong> file.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/551/1*iaOjhH078xGv_M86SyMuHg.png" /><figcaption>Download model poup (Image by Author)</figcaption></figure><p>Unzip the ZIP file, create a model folder in the src directory (src/model) and copy the model.json and the weights.bin into that folder.</p><p>We also need to tell metro to deal with the new file format: *.bin. So please modify the metro.config.js like that:</p><pre>const { getDefaultConfig } = require(&quot;metro-config&quot;);</pre><pre>module.exports = (async () =&gt; {<br>  const {<br>    resolver: { assetExts },<br>  } = await getDefaultConfig();<br>  return {<br>    transformer: {<br>      getTransformOptions: async () =&gt; ({<br>        transform: {<br>          experimentalImportSupport: false,<br>          inlineRequires: false,<br>        },<br>      }),<br>    },<br>    resolver: {<br>      assetExts: [...assetExts, &quot;bin&quot;],<br>    },<br>  };<br>})();</pre><p><strong>Great!</strong> Now that our model is in the project, let’s start using the model to predict the number.</p><h4>Chapter 3: Implement the model, predict and show the result</h4><p>Firstly, we don’t want is to save the photos to our photo library anymore. (unless you want to).</p><p>So comment out that line //await saveToCameraRoll(croppedData.uri);.<br>We also need the base64 string of the cropped image and lastly, we want to pass that base64 string to the EvaluationView via props.</p><p>Let’s modify our CameraView src/screens/CameraView/index.js file again like that:</p><pre>const handlePictureProcessing = async () =&gt; {<br>  const imageData = await takePicture();<br>  const croppedData = await cropPicture(imageData, MASK_DIMENSION);<br>  // await saveToCameraRoll(croppedData.uri);<br>  goToEvaluationView(croppedData);<br>};</pre><pre>const goToEvaluationView = (croppedData) =&gt; {<br>  Navigation.push(props.componentId, {<br>    component: {<br>      name: &quot;evaluationView&quot;,<br>      options: {<br>        topBar: {<br>          title: {<br>            text: &quot;Evaluating ML result&quot;,<br>            color: &quot;white&quot;,<br>          },<br>          background: {<br>            color: &quot;#4d089a&quot;,<br>          },<br>          backButton: {<br>            color: &quot;white&quot;,<br>            showTitle: false,<br>          },<br>        },<br>      },<br>      passProps: {<br>        base64: croppedData.base64 || null,<br>      },<br>    },<br>  });<br>};</pre><p>🏆 <strong>Fantastic job!</strong></p><p>Let’s display the image in the EvaluationView. Import the <strong>Image</strong> from react-native and add the <strong>Image</strong> component to the View container like that</p><pre>&lt;View style={styles.container}&gt;<br>  &lt;Text style={styles.headerText}&gt;ANALYSIS&lt;/Text&gt;<br>  &lt;Image<br>    style={styles.imageContainer}<br>    source={{ uri: `data:image/gif;base64,${props.base64}` }}<br>    resizeMethod=&quot;scale&quot;<br>  /&gt;<br>&lt;/View&gt;;</pre><p>and add the style for the imageContainer underneath the headerText style.</p><pre>imageContainer: {<br> height: 300,<br> width: 300,<br>},</pre><p>The last step is to go to the src/screens/CameraView/helpers.js file and change the saveOptions to base64: true.</p><p>🏆 <strong>Voilà!</strong></p><p>You should see the taken image now in the EvaluationView below the <em>ANALYSIS</em> text.</p><p>Now we want to display the prediction result. We need to add the <strong>Victory Chart</strong> to the EvaluationView together with a few react-native packages</p><pre>import React from &quot;react&quot;;<br>import {<br>  Dimensions,<br>  ActivityIndicator,<br>  SafeAreaView,<br>  View,<br>  Image,<br>  Text,<br>  StatusBar,<br>} from &quot;react-native&quot;;<br>import {<br>  VictoryChart,<br>  VictoryAxis,<br>  VictoryBar,<br>  VictoryTheme,<br>} from &quot;victory-native&quot;;</pre><pre>const { width: DEVICE_WIDTH } = Dimensions.get(&quot;window&quot;);</pre><p>To get the width of the device (needed for the VictoryChart) we use the Dimensions library.</p><p>Then add the <strong>Victory Chart</strong> container. Since we only want to display the chart after we have a prediction result, we add a condition based on the length of the graphData.</p><p>Since we don’t have working model yet we have to add some fake graph data to see the <strong>horizontal bars</strong> of the graph.</p><pre>import React from &quot;react&quot;;<br>import {<br>  Dimensions,<br>  ActivityIndicator,<br>  SafeAreaView,<br>  View,<br>  Image,<br>  Text,<br>  StatusBar,<br>} from &quot;react-native&quot;;<br>import {<br>  VictoryChart,<br>  VictoryAxis,<br>  VictoryBar,<br>  VictoryTheme,<br>} from &quot;victory-native&quot;;</pre><pre>const { width: DEVICE_WIDTH } = Dimensions.get(&quot;window&quot;);</pre><pre>export const EvaluationView = (props) =&gt; {<br>  const graphData = [<br>    { number: 0, prediction: 0.04 },<br>    { number: 1, prediction: 0.02 },<br>    { number: 2, prediction: 0.02 },<br>    { number: 3, prediction: 0.1 },<br>    { number: 4, prediction: 0.85 },<br>    { number: 5, prediction: 0.04 },<br>    { number: 6, prediction: 0.2 },<br>    { number: 7, prediction: 0.12 },<br>    { number: 8, prediction: 0.0 },<br>    { number: 9, prediction: 0.0 },<br>  ];</pre><pre>return (<br>    &lt;React.Fragment&gt;<br>      &lt;StatusBar barStyle=&quot;light-content&quot; /&gt;<br>      &lt;SafeAreaView style={styles.safeArea}&gt;<br>        &lt;View style={styles.container}&gt;<br>          &lt;Text style={styles.headerText}&gt;ANALYSIS&lt;/Text&gt;<br>          &lt;Image<br>            style={styles.imageContainer}<br>            source={{ <br>              uri: `data:image/gif;base64,${props.base64}` <br>            }}<br>            resizeMethod=&quot;scale&quot;/&gt;<br>          &lt;View style={styles.resultContainer}&gt;<br>            {graphData.length ? (<br>              &lt;VictoryChart<br>                width={DEVICE_WIDTH - 20}<br>                padding={{ <br>                  top: 30, bottom: 70, left: 50, right: 30 <br>                }}<br>                theme={VictoryTheme.material}&gt;<br>                &lt;VictoryAxis<br>                  tickValues={[1, 2, 3, 4, 5, 6, 7, 8, 9]}<br>                  tickFormat={[1, 2, 3, 4, 5, 6, 7, 8, 9]}/&gt;<br>                 &lt;VictoryAxis <br>                   dependentAxis <br>                   tickFormat={(tick) =&gt; tick} /&gt;<br>                &lt;VictoryBar<br>                  style={{ data: { fill: &quot;#c43a31&quot; } }}<br>                  barRatio={0.8}<br>                  alignment=&quot;start&quot;<br>                  data={graphData}<br>                  x=&quot;number&quot;<br>                  y=&quot;prediction&quot;/&gt;<br>              &lt;/VictoryChart&gt;<br>            ) : (<br>              &lt;ActivityIndicator size=&quot;large&quot; color=&quot;#4d089a&quot; /&gt;<br>            )}<br>          &lt;/View&gt;<br>        &lt;/View&gt;<br>      &lt;/SafeAreaView&gt;<br>    &lt;/React.Fragment&gt;<br>  );<br>};</pre><p>You should have a screen like that:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/880/1*6BKrfC27yldSi5mkPmQq-g.jpeg" /><figcaption>Chat showing with fake data (Image by Author)</figcaption></figure><p>🏆 <strong>You are a champion!</strong></p><p>We are slowly coming the final part of the tutorial, in where we will load the model and compare the taken photo against the model.</p><p>Please create a util.js in the src directory and paste the following code.</p><pre>/* eslint-disable no-bitwise */<br>/*<br>Copyright (c) 2011, Daniel Guerrero<br>All rights reserved.<br>Redistribution and use in source and binary forms, with or without<br>modification, are permitted provided that the following conditions are met:<br>    * Redistributions of source code must retain the above copyright<br>      notice, this list of conditions and the following disclaimer.<br>    * Redistributions in binary form must reproduce the above copyright<br>      notice, this list of conditions and the following disclaimer in the<br>      documentation and/or other materials provided with the distribution.<br>THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS &quot;AS IS&quot; AND<br>ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED<br>WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE<br>DISCLAIMED. IN NO EVENT SHALL DANIEL GUERRERO BE LIABLE FOR ANY<br>DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES<br>(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;<br>LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND<br>ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT<br>(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS<br>SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.<br> */</pre><pre>/**<br> * Uses the new array typed in javascript to binary base64 encode/decode<br> * at the moment just decodes a binary base64 encoded<br> * into either an ArrayBuffer (decodeArrayBuffer)<br> * or into an Uint8Array (decode)<br> *<br> * References:<br> * <a href="https://developer.mozilla.org/en/JavaScript_typed_arrays/ArrayBuffer">https://developer.mozilla.org/en/JavaScript_typed_arrays/ArrayBuffer</a><br> * <a href="https://developer.mozilla.org/en/JavaScript_typed_arrays/Uint8Array">https://developer.mozilla.org/en/JavaScript_typed_arrays/Uint8Array</a><br> */</pre><pre>export const Base64Binary = {<br>  _keyStr: &quot;ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=&quot;,</pre><pre>/* will return a  Uint8Array type */<br>  decodeArrayBuffer: function (input) {<br>    var bytes = (input.length / 4) * 3;<br>    var ab = new ArrayBuffer(bytes);<br>    this.decode(input, ab);</pre><pre>return ab;<br>  },</pre><pre>removePaddingChars: function (input) {<br>    var lkey = this._keyStr.indexOf(input.charAt(input.length - 1));<br>    if (lkey === 64) {<br>      return input.substring(0, input.length - 1);<br>    }<br>    return input;<br>  },</pre><pre>decode: function (input, arrayBuffer) {<br>    //get last chars to see if are valid<br>    input = this.removePaddingChars(input);<br>    input = this.removePaddingChars(input);</pre><pre>var bytes = parseInt((input.length / 4) * 3, 10);</pre><pre>var uarray;<br>    var chr1, chr2, chr3;<br>    var enc1, enc2, enc3, enc4;<br>    var i = 0;<br>    var j = 0;</pre><pre>if (arrayBuffer) {<br>      uarray = new Uint8Array(arrayBuffer);<br>    } else {<br>      uarray = new Uint8Array(bytes);<br>    }</pre><pre>input = input.replace(/[^A-Za-z0-9\+\/\=]/g, &quot;&quot;);</pre><pre>for (i = 0; i &lt; bytes; i += 3) {<br>      //get the 3 octects in 4 ascii chars<br>      enc1 = this._keyStr.indexOf(input.charAt(j++));<br>      enc2 = this._keyStr.indexOf(input.charAt(j++));<br>      enc3 = this._keyStr.indexOf(input.charAt(j++));<br>      enc4 = this._keyStr.indexOf(input.charAt(j++));</pre><pre>chr1 = (enc1 &lt;&lt; 2) | (enc2 &gt;&gt; 4);<br>      chr2 = ((enc2 &amp; 15) &lt;&lt; 4) | (enc3 &gt;&gt; 2);<br>      chr3 = ((enc3 &amp; 3) &lt;&lt; 6) | enc4;</pre><pre>uarray[i] = chr1;<br>      if (enc3 !== 64) {<br>        uarray[i + 1] = chr2;<br>      }<br>      if (enc4 !== 64) {<br>        uarray[i + 2] = chr3;<br>      }<br>    }</pre><pre>return uarray;<br>  },<br>};</pre><blockquote>Out of respect to the developer please don’t remove the copyright disclaimer <em>😃</em></blockquote><p>Now create another helpers.js file but this time in the EvaluationView directory src/screens/EvaluationView/helpers.js and copy this code</p><pre>import * as tf from &quot;<a href="http://twitter.com/tensorflow/tfjs">@tensorflow/tfjs</a>&quot;;<br>import &quot;<a href="http://twitter.com/tensorflow/tfjs-react-native">@tensorflow/tfjs-react-native</a>&quot;;<br>import { bundleResourceIO, decodeJpeg } from &quot;<a href="http://twitter.com/tensorflow/tfjs-react-native">@tensorflow/tfjs-react-native</a>&quot;;<br>import { Base64Binary } from &quot;../../util&quot;;<br>import { BITMAP_DIMENSION } from &quot;../CameraView/helpers&quot;;</pre><pre>const modelJson = require(&quot;../../model/model.json&quot;);<br>const modelWeights = require(&quot;../../model/weights.bin&quot;);</pre><pre>// 0: channel from JPEG-encoded image<br>// 1: gray scale<br>// 3: RGB image<br>const TENSORFLOW_CHANNEL = 3;</pre><pre>export const getModel = async () =&gt; {<br>  try {<br>    // wait until tensorflow is ready<br>    await tf.ready();<br>    // load the trained model<br>    return await tf.loadLayersModel(bundleResourceIO(modelJson, modelWeights));<br>  } catch (error) {<br>    console.log(&quot;Could not load model&quot;, error);<br>  }<br>};</pre><pre>export const convertBase64ToTensor = async (props) =&gt; {<br>  try {<br>    const uIntArray = Base64Binary.decode(props.base64);<br>    // decode a JPEG-encoded image to a 3D Tensor of dtype<br>    const decodedImage = decodeJpeg(uIntArray, 3);<br>    // reshape Tensor into a 4D array<br>    return decodedImage.reshape([<br>      1,<br>      BITMAP_DIMENSION,<br>      BITMAP_DIMENSION,<br>      TENSORFLOW_CHANNEL,<br>    ]);<br>  } catch (error) {<br>    console.log(&quot;Could not convert base64 string to tesor&quot;, error);<br>  }<br>};</pre><pre>export const startPrediction = async (model, tensor) =&gt; {<br>  try {<br>    // predict against the model<br>    const output = await model.predict(tensor);<br>    // return typed array<br>    return output.dataSync();<br>  } catch (error) {<br>    console.log(&quot;Error predicting from tesor image&quot;, error);<br>  }<br>};</pre><pre>export const populateData = (typedArray) =&gt; {<br>  const predictions = Array.from(typedArray);<br>  return predictions.map((item, index) =&gt; {<br>    return {<br>      number: index,<br>      prediction: item,<br>    };<br>  });<br>};</pre><p>These are our functions to load the model, convert the base64 string to a tensor, predict the digit and populate the data for the victory chart.</p><p>Last but not least, we invoke these functions in the useEffect() hook of the src/screens/EvaluationView/index.js.</p><p>Here the complete code of that view:</p><pre>import React, { useState, useEffect } from &quot;react&quot;;<br>import {<br>  Dimensions,<br>  ActivityIndicator,<br>  SafeAreaView,<br>  View,<br>  Image,<br>  Text,<br>  StatusBar,<br>} from &quot;react-native&quot;;<br>import {<br>  VictoryChart,<br>  VictoryAxis,<br>  VictoryBar,<br>  VictoryTheme,<br>} from &quot;victory-native&quot;;<br>import {<br>  getModel,<br>  convertBase64ToTensor,<br>  startPrediction,<br>  populateData,<br>} from &quot;./helpers&quot;;</pre><pre>const { width: DEVICE_WIDTH } = Dimensions.get(&quot;window&quot;);</pre><pre>export const EvaluationView = (props) =&gt; {<br>  const [graphData, setGraphData] = useState([]);</pre><pre>useEffect(() =&gt; {<br>    const predictDigits = async () =&gt; {<br>      const model = await getModel();<br>      const tensor = await convertBase64ToTensor(props);<br>      const typedArray = await startPrediction(model, tensor);<br>      setGraphData(populateData(typedArray));<br>    };<br>    predictDigits();<br>  }, [props]);</pre><pre>return (<br>    &lt;React.Fragment&gt;<br>      &lt;StatusBar barStyle=&quot;light-content&quot; /&gt;<br>      &lt;SafeAreaView style={styles.safeArea}&gt;<br>        &lt;View style={styles.container}&gt;<br>          &lt;Text style={styles.headerText}&gt;ANALYSIS&lt;/Text&gt;<br>          &lt;Image<br>            style={styles.imageContainer}<br>            source={{ uri: `data:image/gif;base64,${props.base64}` }}<br>            resizeMethod=&quot;scale&quot;<br>          /&gt;<br>          &lt;View style={styles.resultContainer}&gt;<br>            {graphData.length ? (<br>              &lt;VictoryChart<br>                width={DEVICE_WIDTH - 20}<br>                padding={{ top: 30, bottom: 70, left: 50, right: 30 }}<br>                theme={VictoryTheme.material}<br>              &gt;<br>                &lt;VictoryAxis<br>                  tickValues={[1, 2, 3, 4, 5, 6, 7, 8, 9]}<br>                  tickFormat={[1, 2, 3, 4, 5, 6, 7, 8, 9]}<br>                /&gt;<br>                 &lt;VictoryAxis dependentAxis tickFormat={(tick) =&gt; tick} /&gt;<br>                &lt;VictoryBar<br>                  style={{ data: { fill: &quot;#c43a31&quot; } }}<br>                  barRatio={0.8}<br>                  alignment=&quot;start&quot;<br>                  data={graphData}<br>                  x=&quot;number&quot;<br>                  y=&quot;prediction&quot;<br>                /&gt;<br>              &lt;/VictoryChart&gt;<br>            ) : (<br>              &lt;ActivityIndicator size=&quot;large&quot; color=&quot;#4d089a&quot; /&gt;<br>            )}<br>          &lt;/View&gt;<br>        &lt;/View&gt;<br>      &lt;/SafeAreaView&gt;<br>    &lt;/React.Fragment&gt;<br>  );<br>};</pre><pre>const styles = {<br>  safeArea: {<br>    backgroundColor: &quot;#4d089a&quot;,<br>  },<br>  container: {<br>    height: &quot;100%&quot;,<br>    alignItems: &quot;center&quot;,<br>    backgroundColor: &quot;white&quot;,<br>  },<br>  headerText: {<br>    fontSize: 20,<br>    fontWeight: &quot;500&quot;,<br>    color: &quot;#4d089a&quot;,<br>    margin: 20,<br>  },<br>  imageContainer: {<br>    height: 300,<br>    width: 300,<br>  },<br>  resultContainer: {<br>    flex: 1,<br>    justifyContent: &quot;center&quot;,<br>    alignItems: &quot;center&quot;,<br>  },<br>};</pre><p><strong>As I mentioned before, the model will be as good as you train the model.</strong></p><blockquote>In a real-world scenario, a data engineer would use tens of thousands of variations of handwritten digits to train the model. Then would use another set to tweak the model and use a completely new set to check the model performance.</blockquote><blockquote>On a side note before I close this tutorial; if you are a seasoned React Native developer you would have realised by now that with a few manual imports, especially for <em>react-native-unimodules</em>, <em>expo-camera</em> and permission settings, the project will work on Android out of the box too. <em>🤓</em></blockquote><p><strong>I hope you have learnt something new.</strong></p><p>Please leave a comment if I could have done something differently or if you liked this tutorial. After all, we are all here to learn, right? 👨🏼‍🎓</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8089fed59773" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/deep-learning-with-react-native-ios-only-8089fed59773">Deep Learning with React Native (iOS only)</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>
        <item>
            <title><![CDATA[Meet GES Delegate: Thomas Dittmar]]></title>
            <link>https://medium.com/global-entrepreneurship-summit/meet-ges-delegate-thomas-dittmar-2f9f2385a6ac?source=rss-7d69d58a320f------2</link>
            <guid isPermaLink="false">https://medium.com/p/2f9f2385a6ac</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[entrepreneurship]]></category>
            <category><![CDATA[delegate-profiles]]></category>
            <category><![CDATA[ges2016]]></category>
            <dc:creator><![CDATA[Thomas Dittmar]]></dc:creator>
            <pubDate>Fri, 17 Jun 2016 14:18:34 GMT</pubDate>
            <atom:updated>2016-06-17T14:18:34.157Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*l8yepAyLSPLoaDVI8T9bsg.jpeg" /></figure><p>Organization Name: <strong>Bookmemate</strong></p><p>Twitter Handle: <a href="https://twitter.com/Bookmemate"><strong>@bookmemate</strong></a></p><p>Country of Origin: <strong>Australia</strong></p><p>Organization Website: <a href="https://bookmemate.com/"><strong>bookmemate.com</strong></a></p><p><strong>Description of Organization: </strong><a href="https://bookmemate.com/">Bookmemate</a> is an easy to use and very affordable business management cloud solution. Users can create and send quotes in minutes, manage &amp; book their clients, display clients history, easily remind clients of appointments by sending out SMS’s, record expenses and extra income and produce a profit and loss statement from within the app with a click of a button. This sets <a href="https://bookmemate.com/">Bookmemate</a> completely apart from any other system.</p><p><strong>What inspired you to start this organization:</strong> When I started <a href="https://bookmemate.com/">Bookmemate</a> I had a vision: to help small or medium sized businesses better manage their bookings and finances. Most of them were running a business with an appointment book, a spreadsheet and email, which had several disadvantages such as finding past bookings, calculating viability or produce financial documents for the business.</p><p><strong>What is the next big step you hope to help your organization reach?</strong><a href="https://bookmemate.com/">Bookmemate</a> has long passed acceptance testing and many happy customers around the world are enjoying the software. We are now ready to scale. Hence, the next logical step is to find partners or investors in the US or elsewhere that can take the product to the next level.</p><p><strong>What has been your biggest obstacle as an entrepreneur? </strong>I think the biggest obstacle for me is to find the right partner; someone who is in line with my vision.</p><p><strong>What advice would you give other emerging entrepreneurs ? “</strong> As a startup CEO, I slept like a baby. I woke up every 2 hours and cried.” — Ben Horowitz</p><p>My advice is to not give up and stay focused. Remember, it took the guys from Rovio over 8 years to land the “overnight success” Angry Bird.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2f9f2385a6ac" width="1" height="1" alt=""><hr><p><a href="https://medium.com/global-entrepreneurship-summit/meet-ges-delegate-thomas-dittmar-2f9f2385a6ac">Meet GES Delegate: Thomas Dittmar</a> was originally published in <a href="https://medium.com/global-entrepreneurship-summit">Global Entrepreneurship Summit</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How You Can Effortlessly Manage Your Small Business in Less than an Hour.]]></title>
            <link>https://medium.com/@bookmemate/how-you-can-effortlessly-manage-your-small-business-in-less-than-an-hour-b10a4fc0d256?source=rss-7d69d58a320f------2</link>
            <guid isPermaLink="false">https://medium.com/p/b10a4fc0d256</guid>
            <category><![CDATA[small-business]]></category>
            <category><![CDATA[appointment-scheduling]]></category>
            <category><![CDATA[financial-services]]></category>
            <category><![CDATA[saas-tools]]></category>
            <category><![CDATA[bookmemate]]></category>
            <dc:creator><![CDATA[Thomas Dittmar]]></dc:creator>
            <pubDate>Fri, 05 Feb 2016 07:54:21 GMT</pubDate>
            <atom:updated>2016-06-16T06:41:39.246Z</atom:updated>
            <content:encoded><![CDATA[<p>When I started <a href="https://bookmemate.com/">Bookmemate</a> I had a vision: to help small or medium sized businesses better manage their bookings and finances. I spoke with a variety of business owners: plumbers, electricians, beauticians and more, to find out what method they most commonly used to keep their books.</p><p>The result of this research gave me insight into the challenges of running a small to medium business. They were using paper books for their bookings, and most weren’t managing their finances at all.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*IlqDdcs7KtOKluuI9mfGIQ.jpeg" /><figcaption>An error-prone paper appointment book.</figcaption></figure><p>Running a business with a notebook and pen severely disadvantages a business. Finding past bookings among the clutter of paper book bookings is very difficult. It is also an extremely messy and impractical process to calculate revenue or produce financial documents for the business.</p><p>After speaking to these businesses and looking over their finances, I was shocked to learn many of them were not even viable; they did not know their financial situation, and could not populate a profit and loss statement.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/480/1*Dp44MpRwAfwzXKKw00ovUQ.jpeg" /><figcaption>Calculating Profit &amp; Loss by hand is tedious.</figcaption></figure><p>This is a reality for many small to medium sized businesses, and tragically will send them out of business very quickly. All businesses should have systems to manage their bookings and finances, regardless of their size. An affordable and easy to use system is greatly needed.</p><p><a href="https://bookmemate.com/">Bookmemate</a> is a simple, easy to use, affordable all-in-one booking and financial system. Users can easily remind clients of appointments by sending out texts through the app, as well as schedule appointments. The task management function means completed tasks remain in the system, as do client information, tasks completed and revenue received, for quick reference later.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*IMo_7N0nuuyR3cRB9KiRKA.png" /></figure><p>By saving revenue information as well as recording expenses, such as rent, gas, etc. The app even manages the inventory. The user can then produce a profit and loss statement from the app. This sets <a href="https://bookmemate.com/">Bookmemate</a> completely apart from any other system.</p><p>Business owners can send a client a quote in under a minute, in a professional format, via email. Once the quote has been accepted, you can send an invoice or receipt to the client.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/746/1*dMCN4dGJk6PuL3bzTwjEmg.png" /></figure><p><a href="https://bookmemate.com/">Bookmemate</a> is a simple, easy to use and affordable record keeping system that you need for your small or medium sized business. <a href="https://bookmemate.com/">Bookmemate</a> will allow businesses that would not otherwise be able to afford a record keeping system to flourish. It’s a low cost, high quality system that will save their business in the long run; it is an app that helps small businesses meet their true potential.</p><p><a href="https://bookmemate.com/">Bookmemate</a> is packed with features like:</p><p>● Appointment management</p><p>● SMS reminders</p><p>● Client management</p><p>● Income &amp; Expense management</p><p>● Quote &amp; Invoice management</p><p>● Inventory management</p><p>● Sending &amp; receiving SMS</p><p>● Automated receipts</p><p>● Client statistics &amp; history</p><p>● Expense revenue charts</p><p>● Financial statements like profit &amp; loss</p><p>● Google Calendar sync and many more</p><p>Thomas,</p><p>Founder of Bookmemate</p><p><a href="https://bookmemate.com/#testimonial%29">https://bookmemate.com</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/164/1*2ry3uGiwDr3vu10K7U57_g.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b10a4fc0d256" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>