<?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 Noman Mazher on Medium]]></title>
        <description><![CDATA[Stories by Noman Mazher on Medium]]></description>
        <link>https://medium.com/@noman.mazher?source=rss-2936a2f6f4d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*6_8t-otGuJCxUSNxDmLZeg.jpeg</url>
            <title>Stories by Noman Mazher on Medium</title>
            <link>https://medium.com/@noman.mazher?source=rss-2936a2f6f4d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 09:02:04 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@noman.mazher/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[How to Build a Task-Specific AI Agent with No-Code Tools (Make.com + n8n)]]></title>
            <link>https://medium.com/@noman.mazher/how-to-build-a-task-specific-ai-agent-with-no-code-tools-make-com-n8n-2e00e70908d9?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/2e00e70908d9</guid>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Sat, 26 Apr 2025 07:46:34 GMT</pubDate>
            <atom:updated>2025-04-26T07:46:34.828Z</atom:updated>
            <content:encoded><![CDATA[<p>How to Build a Task-Specific AI Agent with No-Code Tools (Make.com + n8n)</p><p>AI agents are no longer science fiction — they are everyday tools powering real businesses. But many people believe you need deep coding skills to build one. That’s simply not true.</p><p>Thanks to powerful no-code platforms like <strong>Make.com</strong> and <strong>n8n</strong>, anyone can design, deploy, and manage task-specific AI agents without writing a single line of code.<br> In this article, I’ll show you how to create a smart, automated Research Assistant agent that fetches real-time data, processes it using GPT, and delivers valuable summaries — all without touching Python.</p><p>If you’re a solo founder, freelancer, or tech-savvy entrepreneur, this guide is for you.</p><h3>What Is a Task-Specific AI Agent?</h3><p>At a high level, an <strong>AI agent</strong> is a system that:</p><ul><li><strong>Thinks</strong> (analyzes a situation)</li><li><strong>Plans</strong> (figures out the next best action)</li><li><strong>Acts</strong> (executes tasks autonomously)</li></ul><p>A <strong>task-specific</strong> agent is tightly focused on one type of problem (e.g., summarizing news articles, monitoring competitor prices, generating leads). It’s not meant to “do everything” — it’s specialized and efficient.</p><p>Building narrow agents has <em>huge advantages</em>:</p><ul><li>Faster development</li><li>Cheaper to run</li><li>Easier to monitor and improve</li></ul><p>Now, let’s get into the fun part: actually building one.</p><h3>The Use Case:</h3><p><strong>Goal:</strong> Build an AI agent that:</p><ol><li>Pulls the latest articles from a news API (e.g., NewsAPI.org)</li><li>Sends article content to GPT-4</li><li>Returns a 3-bullet-point summary</li><li>Delivers it to your Slack or Email</li></ol><p>Perfect for anyone who needs daily briefings without sifting through endless feeds!</p><h3>Step-by-Step Guide</h3><h3>Step 1: Set Up Your No-Code Platform</h3><p>You can use either:</p><ul><li><strong>n8n</strong>: Open-source, self-hosted (or cloud-hosted) tool.</li><li><strong>Make.com</strong>: Hosted SaaS platform, highly visual.</li></ul><p>Both allow you to build powerful workflows by connecting APIs, databases, LLMs, and other services.</p><blockquote><strong><em>Tip:</em></strong><em> For beginners, Make.com might feel more intuitive. For customization nerds, n8n offers more control.</em></blockquote><h3>Step 2: Connect to the News API</h3><ul><li>Create an account at <a href="https://newsapi.org/">NewsAPI.org</a>.</li><li>Grab your <strong>API key</strong>.</li><li>In Make.com or n8n:</li><li>Add an <strong>HTTP Request</strong> module/node.</li><li>Set the URL to https://newsapi.org/v2/top-headlines?country=us&amp;apiKey=YOUR_KEY</li><li>Choose GET method.</li><li>Test the request and see if you get article titles and content!</li></ul><p><strong>Output example:</strong></p><pre>json<br>{<br>  &quot;articles&quot;: [<br>    {<br>      &quot;title&quot;: &quot;OpenAI releases GPT-5 preview&quot;,<br>      &quot;description&quot;: &quot;New advancements in AI models...&quot;,<br>      &quot;url&quot;: &quot;https://example.com/article&quot;<br>    }<br>  ]<br>}</pre><h3>Step 3: Add GPT-4 for Summarization</h3><ul><li>Add an <strong>OpenAI GPT module</strong>.</li><li>Set the prompt to something like:</li></ul><pre>Summarize the following news article in 3 bullet points:<br>{article_content}</pre><blockquote><strong><em>Pro Tip:</em></strong><em> You can even craft prompts to change tone: “Make it sound formal”, “Make it exciting”, “Add emojis”, etc.</em></blockquote><h3>Step 4: Deliver the Output</h3><ul><li>Add an <strong>Email</strong> module, <strong>Slack webhook</strong>, or <strong>Telegram Bot</strong>.</li><li>Map the GPT-4 output into your message body.</li><li>Set the automation to run <strong>daily</strong>, or <strong>on trigger</strong>.</li></ul><h3>Step 5: (Optional) Add Memory</h3><ul><li>Want your agent to <em>remember</em> past summaries?</li><li>Add a <strong>Data Store</strong> (in Make.com) or <strong>Postgres/MongoDB node</strong> (in n8n).</li><li>Store previous articles and compare new results to avoid duplicates.</li></ul><h3>Visual: Example Agent Workflow</h3><pre>[Trigger Timer] --&gt; [Fetch Articles] --&gt; [Summarize with GPT-4] --&gt; [Send to Slack]</pre><pre>[Trigger] <br>  ↓<br>[News API Call]<br>  ↓<br>[Check for new articles]<br>  ↓<br>[Summarize new articles with GPT-4]<br>  ↓<br>[Save to Database]<br>  ↓<br>[Send to User]</pre><h3>Scaling Ideas</h3><p>Once you build one agent, you can expand it:</p><ul><li>Add <strong>multi-channel delivery</strong> (Slack + Email + SMS)</li><li>Build a <strong>dashboard</strong> showing summaries</li><li>Combine <strong>different data sources</strong> (e.g., Reddit + Twitter monitoring)</li><li>Let users <strong>ask follow-up questions</strong> via a chatbot!</li></ul><h3>Final Thoughts</h3><p>You don’t need to be a software engineer to build smart AI agents.<br> With Make.com, n8n, and clever use of GPT models, you can automate complex tasks, create useful services, and even build businesses around these tiny agents.</p><p>This is just the beginning — imagine what happens when you layer multiple agents together!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2e00e70908d9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to add remote ID to an Ardupilot]]></title>
            <link>https://medium.com/@noman.mazher/how-to-add-remote-id-to-an-ardupilot-40d3d1a4d061?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/40d3d1a4d061</guid>
            <category><![CDATA[ardupilot]]></category>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Sun, 08 Jan 2023 17:05:01 GMT</pubDate>
            <atom:updated>2023-01-08T17:05:01.332Z</atom:updated>
            <content:encoded><![CDATA[<p>To an Archer pilot autopilot it assumes that you already have set up a build<br>environment and are familiar with how to build firmware and archipilot in general for this video we’ll be using a cube orange autopilot and esp32s3 board for the Archer remote and a can external transceiver to convert it<br>to a drone can remote ID the first step is to create archipilot firmware that has open drone ID capabilities and install that then we’ll need to program an Archer remote ID module or remote ID module with its necessary firmware and parameters then set up those RG remote ID parameters in the remote ID module and finally discuss the way a user would interface<br>and configure the system using a ground control station so to build the firmware we’ll create a separate branch and create three new HW I mean two new HW def files and a new params def based off of the cube orange<br>we’ll call the branch Cube Orange odid and the hardware hwdef.dat file instead of the board ID of 140 for cube orange we’ll put a 1040 10 140 and enable an open drone ID capability in the firmware we’ll do the same thing for the bootloader HW def so it has a matching board ID and open drone capabilities and and then very specific some uh open drone ID parameters will make them read only these are set up as shown for drone can versus a mavlink connection and then we’ll go and create a bootloader based off of those files Cube orange Dash odid and it’ll create and put a bootloader in the bootloader directory the next thing we do is configure to build the firmware and we’ll do a Waf configure for the open drone ID version of cube Orange and then we’re going to build plane firmware for it we’ll off plane and I’m showing uh the automatic upload now this upload just to show you will fail because the firmware we’re creating has a board date ID of 10140 and the cube orange with its standard existing firmware has a board ID of 140. so this operation of the final operation of loading the firmware will actually fail but I just wanted to show that it will fail<br>the firmware is created and it’s attempting to load it onto the board as<br>soon as the board goes into bootloader mode there’ll be a discrepancy as you see of the board IDs therefore we have to force load the software using the uploader.py tool to overcome that board ID check for the load so I’m using uploader.py with force and I’m pointing to the firmware that we created in the build step previously now it is loading it because we’re<br>forcing it to load you see the board ID discrepancies there the current boot loader is 140 the firmware is 10 140 so it resists loading that but we’re forcing it now when you build open drone ID firmware automatically it puts in the checks such that when the bootloader is finally installed<br>with a board ID of 10 140 it will not ever run any firmware that has been<br>Force overloaded into the main firmware area so now we need to update the bootloader from the firmware that is currently running to replace that old boot boot loader that’s right for standard firmware you can see we’re running the open did main firmware and now we’re going to send the command in map proxy flash bootloader and you see it’s flashed the new bootloader so now that we’ve built and loaded and installed the firmware on the autopilot let’s program up the remote ID module this with RG<br>remote ID so we’ll plug into its USB plug programming port on the module and now we’ll download firmware we go to the archipilot remote ID<br>GitHub get the latest release and save it this is the S3 board so we’ll download that then the next thing we’ve got to do is<br>Flash program it with the expressive firmware programmer that you can<br>download from the documentation you can get links to that we’ll open up the programmer select the board S3 select the USB programming port on that board then select the firmware that we just downloaded for the Estuary board select it it’s on com Port 15 already I’ve set that up<br>and now it’s programming so now the remote ID module has been<br>programmed we will probably need to set up some parameters such as the uas ID type so forth and so on as well as perhaps changing capabilities or adding security features so we disconnect and connect the USB to the flight controller connect to the flight controller and you can see it’s running the open did firmware and it hasn’t been all set up yet either from the manufacturer side or the user side so it’s giving pre-arm warnings so we’ll go into setup into the optional Hardware drone can and this is on the second can bus so I click the MAV link for can two you see the remote ID module click the menu get to its parameters and now we can see all the parameters for that are in the remote ID module particularly you will have to change the uas id id type and uh those fields to get that’s what’s proper for your manufacturing as well as any capabilities you may change you can also add private public key pairs for security which we’ll discuss later on as<br>well as setting up locking levels to prevent parameter or firmware updates<br>via the USB at various levels we’ll discuss those later on in the video so now we’re going to discuss what an end user of the system must do to enable<br>it connects to the autopilot equipped with a remote ID module and there are several things that must be put in place before he can arm and fly the vehicle so if he attempts to just arm the vehicle and we’ll use Force arming here since we have other things as pre-arm failures normally you can force arm over anything however you cannot force arm over open eye degree open a drone ID issues or remote ID issues as you see here it will not even allow Force Army the only way to arm the vehicle even Force arm the<br>vehicle is to satisfy the requirements the first is you’ve got to have position<br>information into the ground control station either by a GPS attached<br>serially or you could use an Android phone using this app GPS to Bluetooth<br>over Bluetooth connection to the Ground Control station to provide osition<br>information and that’s what we’ll do in this case Simple app you just start it<br>if you have a Bluetooth com Port you attach connect through that com port<br>and you’ll be getting GPS position information into the ground control station as you see here that’s one requirement but it’s not the only requirement if someone attempts to arm the vehicle with just the GCS attached but failing to enter in some more important information<br>again you can’t even Force arm the vehicle in some jurisdictions you’ll be required to enter in the operator ID it doesn’t prevent arming currently but it’s required in certain areas now this should allow arming if all the pre-arm checks are gone or forced arming if all the other pre-arm checks are not resolved however there’s still a bug and the communication between the module and the firmware and even though there is already built into<br>the Archer pilot modules it’s uas ID it’s type and UA type you have to enter these in to allow it to arm this is expected to be resolved and I’ll actually probably remove this section of the video Once that’s resolved but for right now you still have to do that you see you’ve got a green light on remote ID okay if you’re looking at the pre-arm checks they’ve been throwing uh amongst the other pre-arm checks the remote ID or pre-arm check and every 30 seconds or so you get a set of pre-armed messages you’ll notice the next set does not have that UA type requirement in it okay now it’s able to be armed if you resolve the other pre-arms or Force arm if you don’t resolve the other pre-arms well Force Armament and you see it is indeed armed<br>so now you can see the remote ID module information presented on one of the several phone apps open drone ID us like the omcs Drone scanner hatever and you see the location of both the operator and the vehicle<br>on that application as well as if you click on the info you will see all the details you’ll notice that the uas ID abcd123456789 comes from the remote ID module the operator ID came from Archer pilot tester what I had put in for the operator ID the location of the vehicle as well as the operator location.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=40d3d1a4d061" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What to look at when buying a used car]]></title>
            <link>https://medium.com/@noman.mazher/what-to-look-at-when-buying-a-used-car-75f455e697e8?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/75f455e697e8</guid>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Sun, 08 May 2022 08:43:22 GMT</pubDate>
            <atom:updated>2022-05-08T08:43:22.015Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*MfE-KwsUD04VuH_ojw_CEw.jpeg" /></figure><p>okay, guys I just want to go over what to look for when you’re buying a car used off of maybe side of the road; you want a cheap vehicle for a couple of grand, a few things you want to keep in mind</p><h4>1. Don&#39;t buy it at night</h4><p>The first thing you want to do is don’t buy it at night take a look at it in the daylight you want to see it in good sunlight and you want to be able to compare panels to each other to make sure that they all look uniform if it’s been painted first warning signs stay away it’s either been in an accident or it’s rusty if one panel rusty most likely other panels rusty and it’s on its way out.</p><h4>2. Check the tires</h4><p>Tires are very important if you’re spending a few grand on a vehicle, even eight grand if the tires of war you’re looking at could possibly another eight hundred thousand for new boots, so you want to look at the tires you won&#39;t turn them all the way out one way, and you want to run your feet, your hands over and make sure that it’s wearing evenly if it’s feathered at all where it’s cupped it means it’s a bad alignment and that could be worn parts, and you’d have to replace those suspension parts or again it could be in an accident so make sure that your goods are in good shape underneath the hood first and foremost</p><h4>3. Check the transmission fluid</h4><p>you can want to check your transmission fluid. Usually a yellow handle, you want to look for the nice red fluid that’s not low, and you want to smell it and make sure that it’s not burnt. You want to check your engine oil to make sure that it’s not low by any means and that it’s in good shape, nice and clear. Again doesn’t smell good oil isn’t as telltale because there was an issue they would have.</p><h4>4. Check the engine oil and brake oil</h4><p>changed the oil beforehand, and you can actually tell them how much you break of war, but how much fluid is left in your reservoir is the brake fluid reservoir. If your brakes are wor, that means that your Pistons are coming out farther, and this reservoir will be lower if it’s full. It’s a good sign that your brakes are in good shape, but somebody could have topped up the fluid as well, so what I’m saying is low oil means that the brakes are worn when the reservoir is full; it means it could be could have really good brakes or somebody’s topped it is off, so we have to take that with a grain of salt</p><h4>5. Check the belts corrosion, rust</h4><p>you want to check over the belt&#39;s general condition of tt you want to check to make sure there’s a little corrosion and rust on the exhaust and aluminum parts.</p><h4>6. check your antifreeze by OBD2 reader</h4><p>when it’s cold, make sure that there’s no oil or anything in there, and it looks nice and green or pink depending on what’s in there now obd2 reader, you can pick these up really cheap now. This one I paid forty-five bucks for, and every single car built after 99 has an overlay on the console, turn your ignition on and best case you want to look for zero diagnostic out of codes, you can read the VIN, you can read the codes, and so just hit enter it checks your can bus which is your computer for anything logged and hopefully it will come up no diagnostic.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=75f455e697e8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to prevent ATV from overheating]]></title>
            <link>https://medium.com/@noman.mazher/how-to-prevent-atv-from-overheating-608a4225aecf?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/608a4225aecf</guid>
            <category><![CDATA[atv]]></category>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Wed, 20 Apr 2022 10:46:09 GMT</pubDate>
            <atom:updated>2022-04-20T10:46:09.761Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lA7GVprKjLWWhG0_vHHLoA.jpeg" /></figure><p>Hi folks,</p><p>Summer is just out of the door; it is time to hit the trails with your ATV. The most panic issue for ATVs in summer is overheating. Be active, webrings some precautionary measures and some solutions to prevent your quad from overheating.</p><h4>How you can say your ATV is over-heated</h4><p>This is probably the first question you have to ask yourself since overheating is not always looking warm. You’ve been tearing up the terrain with your ATV in the sun, and your ATV feels hot. But is this the cause of your surroundings, or is your ATV’s engine actually overheating?</p><p>There are plenty of signs you should look for, including the following:</p><ul><li>Your vehicle loses power</li><li>Fluid is dripping from your vehicle</li><li>Your vehicle’s case is particularly hot</li><li>The engine is making a ticking noise, indicating oil burning off</li><li>A strange smell is coming from the coolant</li><li>Steam is visible</li><li>The piston is expanding</li></ul><p>If you see any of these signs, you should be cautious, as continued use upon overheating is dangerous.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*q9yvrK7KPeZviZAzD1brNw.jpeg" /></figure><h3>How does overheating damage your ATV</h3><p>In terms of your vehicle’s well-being, overheating may cause it to stop functioning correctly at the moment — and it may leave long-term or even permanent damage to your vehicle.</p><p>The coolant needs a break in order to work effectively and keep your vehicle cool. The engine also might lose its lubricant upon being overheated.</p><p>It is also unsafe to ride a vehicle that is overheating.</p><h3>How to stop ATV from overheating</h3><p>What can you do if you’re on the trail when you notice an overheating problem? Don’t worry — there are plenty of strategies you can implement to keep your four-wheeler riding at the right temperature.</p><p>First, remember that prevention is key. Whenever possible, keep your radiator and its screen clean. Also, make sure to check often whether your coolant needs to be refreshed or not. You can refer to your vehicle’s manual, or the manufacturer’s recommendations, to know how often you should change the fluid, though usually, it is every few years.</p><p>There are some upgrades you can get to protect your vehicle against overheating. Electrical radiator fans are one popular option</p><h3>CHECKLIST: What if my vehicle does not start at-allSUMMER ATV STORAGE TIPS</h3><p>If your vehicle isn’t stored properly during the summer, its ability to keep cool might wear off. Keeping it in a cool, dry place (like a shed or a garage) is the best way to store it — even in the summer.</p><p>Since that’s not always an option, take caution if you store your ATV outside. We recommend using a heavy-duty ATV cover and chaining it to something for safety’s sake. Using a battery tender will protect your battery’s lifespan. When storing your vehicle outside, regularly check its fuel levels and its tires.</p><h4>CONCLUSION</h4><p>Dealing with an ATV that’s too hot is a frustrating experience for any rider. That’s why prevention is key — and we hope the tips above help you solve any temperature-related calamity you experience.</p><p>It’s possible that your vehicle is so overheated that the engine simply won’t budge. There are plenty of reasons why this might happen, with an overheated engine being one of them. Otherwise, run down this list of potential problems to ensure that your ATV is ready for the trail:</p><ul><li>Make sure the battery is charged and installed correctly</li><li>Ensure that the spark plug is working and installed correctly</li><li>Check to see if you have enough fuel, or that fuel is leaking</li><li>Clean the air filter to ensure the ATV is getting enough airflow</li><li>Make sure the engine’s individual parts are working; this may include the head gasket, cylinder walls, or piston ring</li><li>Make sure the fuel in the vehicle isn’t too old</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=608a4225aecf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[“Intelligent Transport System,” some common terminology]]></title>
            <link>https://medium.com/@noman.mazher/intelligent-transport-system-some-common-terminology-63ae4972be23?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/63ae4972be23</guid>
            <category><![CDATA[v2x]]></category>
            <category><![CDATA[intelligent-transport]]></category>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Sat, 16 Apr 2022 11:22:01 GMT</pubDate>
            <atom:updated>2022-04-16T11:26:19.345Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pkN37MaPmbMdDrr-KuxP9g.jpeg" /></figure><p>Intelligent Transport System (ITS) is one of the most interesting taking technology covering IT, electrical, transportation and telecommunication research and development areas.</p><p>This article will introduce some common terminology used in ITS. By reading these terminologies, users will be able to read and grasp more technical writing on ITS.</p><h4><strong>Cooperative Intelligent Transport System (C-ITS)</strong></h4><p>C-ITS is, also known as connected vehicle technology uses wireless technologies to enable real-time communication between vehicles moving along roads, traffic signals, infrastructure and other mobile devices</p><h4><strong>V2X communication:</strong></h4><p>Vehicular applications and communications technologies are collectively referred to as <strong>vehicle to everything (V2X</strong>), which is broadly classified into four different types: vehicle to vehicle (V2V), vehicle to infrastructure (V2I), vehicle to the network (V2N), and vehicle to pedestrian (V2P)</p><h4>DSRC(dedicated Short-Range Communication) :</h4><p>For physical communications, there are two main types of communications in the modern V2X system: short-range communications and long-range communications. For short-range communications, the U.S. government released the Dedicated Short-Range Communications (DSRC) standards, while the European Telecommunications Standards Institute<br>(ETSI) released Intelligent Transportation System (ITS)-G5 standards. Both standards are based on IEEE 802.11p.</p><h4><strong>Long-Term Evolution</strong> (<strong>LTE</strong>):</h4><p>Long-Term Evolution (LTE) is a standard for wireless broadband communication for mobile devices and data terminals. A recently widely deployed LTE network has been envisioned to support V2X communication in 3GPP release 14.</p><h4>RSU ( Roadside unit):</h4><p>A Roadside Unit is a DSRC transceiver mounted along a road or pedestrian passageway. An RSU may also be mounted on a vehicle or is hand-carried, but it may only operate when the vehicle or hand-carried unit is stationary.</p><h4>OBU (On-Board Unit) :</h4><p>OBU is a communication device mounted on vehicles. It allows DSRC communications with other OBUs or RSUs.</p><h4>UE (User Equipment):</h4><p>UE is a device that is used by end-user to communicate with eNodeB and other UEs</p><h4>eNobeB :</h4><p>it is hardware that is connected to mobile phone network communications directly wirelessly with mobile handsets</p><h4>CAM( Cooperative Aware Message):</h4><p>are distributed within the ITS-G5 (802.11p) network containing the vehicle,s geographical position, speed and heading to the neighboring vehicles</p><h4>Decentralized Environmental Notification Message(DENM):</h4><p>It is sent by any vehicle or RSU. it describes any specific events like emergency brake or accident etc</p><h4>Collective Perception Message (CPM):</h4><p>describe any object such as an obstacle etc</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=63ae4972be23" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Electronic Self Label for grocery stores]]></title>
            <link>https://medium.com/@noman.mazher/electronic-self-label-for-grocery-stores-19242b560eb9?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/19242b560eb9</guid>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Wed, 09 Mar 2022 04:58:43 GMT</pubDate>
            <atom:updated>2022-03-09T04:58:43.336Z</atom:updated>
            <content:encoded><![CDATA[<p>The idea of electronic shelf labels (ESLs) is not a new one, and the application of the concept has become fairly common in Europe. But U.S. grocers have been slow to the uptake for a variety of reasons, including a perceived low return on investment. This, however, may be about to change as ESLs evolve to perform tasks that go beyond a quicker price change.</p><p>ESLs are gaining new efficiencies that can cut down labor costs, manage stock, streamline e-commerce, and communicate with shoppers in new ways.</p><p>Brian Salzman, national accounts manager, West, for ESL provider <a href="https://aperionsolutions.com/">Aperion Solutions</a>, a subsidiary of Bridgeton, Mo.-based Hussmann Corp., said grocery stores are now looking at anywhere between an 18-month and two-year payback if they choose the correct functionalities. Given these advances, Salzman said ESL companies are “on the cusp of breaking this thing open,” and predicts that this will be the year in which “ESLs grab hold with retailers because there are many more things to build into the profitability and there’s much more value than ESLs have delivered in the past.”</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PaiG3y60uIsCEzD_DrOuOQ.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=19242b560eb9" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A beginner guide for convolutional neural network]]></title>
            <link>https://medium.com/@noman.mazher/a-beginner-guide-for-convolutional-neural-network-aaa315996879?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/aaa315996879</guid>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Thu, 03 Mar 2022 09:49:25 GMT</pubDate>
            <atom:updated>2022-03-03T09:49:25.616Z</atom:updated>
            <content:encoded><![CDATA[<h3>A convolutional neural network is a branch of deep learning mainly developed for image data use. Yann LeKunn, a postdoctoral researcher, first introduced CNN. At that time algorithm developed by LeKunn was famous by LeNet by the name of LeKnn. In 2012, AlexNet presented in his research the new dimension of deep learning and added CNN as an essential part of deep learning and computer vision.</h3><p><strong>Components of CNN</strong></p><p>The working of CNN is based on the following two main components</p><p>1. Feature extraction</p><p>2. Classification</p><p><strong>Feature extraction: </strong>Different layers are applied to an image, and its features are extracted in the feature extraction phase. That features are in turn sent for classification.</p><p><strong>Classification: </strong>The extracted feature of the image is classified in this stage. This classification happens based on the problem s target variables.</p><h3><strong>Layers of CNN</strong></h3><p>A typical CNN model exists of following layers</p><ol><li>Input layer</li><li>Convolution layer + Activation function</li><li>Pooling layer</li><li>Fully Connected Layer</li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*GA_uqgUvfoBKSYlQ1yuDEQ.jpeg" /></figure><h3><strong>Input layer:</strong></h3><p>Image is extracted from the source in the input layer. Image can be either RGB or Grayscale. Every image in a computer consists of pixels that range from 0 to 255. We have to convert the ranges of each pixel from 0 to 1 before passing it into the next layer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*nPEc8AbDV1N_eNVaTQBfQw.png" /></figure><p>This figure shows an image of 4 x 4 size and three color channels.</p><h3><strong>Convolution Layer:</strong></h3><p>We apply a filter to our input image in the convolution layer for detecting its features. These filters can be used multiple times and create a feature map. This feature mapping help in classifying our target image.</p><p>In this example, we take a 2D input image for convenience. The figures mentioned above show an input image of 6 x 6 size. We applied a filter of 3 x 3 for detecting features. Although in this example, we applied only one filter yet, there can be more than one folder that can be applied for feature extraction</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/897/1*a0f4aeS3xBxGf3T93PRbZA.png" /></figure><p>The figure below shows the result after applying the filter; the resultant map is size 4 x 4, containing information about the image.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/868/1*BJR1-FUxyaj765CdVQujRQ.png" /></figure><p>To understand this process, let see the green part of the map. The pixel values are multiplied according to the importance of the filter. After multiplication, the cell values are summed up. This calculation is also shown in the figure.</p><p>In the below figure, this filter is shifted by one column. This process of jumping to the following columns and rows is called astride. In this example, we tool astride as 1, so we shift only 1 column</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/865/1*QWU_9v7fvAQGGrKXlr12Hg.png" /></figure><p>The filters pass complete images to get the final feature map from the same process. After getting the feature map, the next step is to apply the activation function to introduce nonlinearity. One point of notice is that feature maps are smaller than the size of the original image. As much as we increase the stride value, the feature map’s size will decrease.</p><h3>Pooling Layer:</h3><p>The pooling layer comes after the convolutional layer. The purpose of the pooling layer is to reduce the dimension of feature maps to reduce computational time. The pooling layer produces the output of a lower resolution image, yet this image contains all critical information of the source image. Two common types of pooling available are Average pooling and max pooling. Here we will show you how the Max pooling works.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/769/1*jyP-G2TqgkD9w5toShckUg.png" /></figure><p>This image shows the working of max pooling. In this example, we took a pooling layer of size 2 x 2 and a stride of 2.</p><p>Fully connected layer:</p><p>The first two earlier described layers worked for feature extraction while the fully connected layer functions as a classification part of CNN. Simply speaking fully connected layer classifies the input image into a label. This layer makes the connection from the previous layer, get extracted features, to the output layer. The below-mentioned figures show the complete process of CNN.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c2ubkAfmplp3r7hTz-mk5A.png" /></figure><h3>Implementation of CNN in python</h3><p>We use the MNIST(Modified National Institute of Standards and Technology) data set in this example. MNIST consists of 60,000 images, each of 28 x 28 grayscale images.</p><h3><strong>Importing of required libraries:</strong></h3><blockquote>#importing the required libraries</blockquote><blockquote>from tensorflow.keras.datasets import mnist</blockquote><blockquote>from tensorflow.keras.models import Sequential</blockquote><blockquote>from tensorflow.keras.layers import Conv2D</blockquote><blockquote>from tensorflow.keras.layers import MaxPool2D</blockquote><blockquote>from tensorflow.keras.layers import Flatten</blockquote><blockquote>from tensorflow.keras.layers import Dropout</blockquote><blockquote>from tensorflow.keras.layers import Dense</blockquote><h3><strong>Loading and reshaping data:</strong></h3><blockquote>#loading data</blockquote><blockquote>(X_train,y_train) , (X_test,y_test)=mnist.load_data()</blockquote><blockquote>#reshaping data</blockquote><blockquote>X_train = X_train.reshape((X_train.shape[0], X_train.shape[1], X_train.shape[2], 1))</blockquote><blockquote>X_test = X_test.reshape((X_test.shape[0],X_test.shape[1],X_test.shape[2],1))</blockquote><blockquote>#checking the shape after reshaping</blockquote><blockquote>print(X_train.shape)</blockquote><blockquote>print(X_test.shape)</blockquote><blockquote>#normalizing the pixel values</blockquote><blockquote>X_train=X_train/255</blockquote><blockquote>X_test=X_test/255</blockquote><h3><strong>working on CNN model:</strong></h3><blockquote>#defining model</blockquote><blockquote>model=Sequential()</blockquote><blockquote>#adding convolution layer</blockquote><blockquote>model.add(Conv2D(32,(3,3),activation=’relu’,input_shape=(28,28,1)))</blockquote><blockquote>#adding pooling layer</blockquote><blockquote>model.add(MaxPool2D(2,2))</blockquote><blockquote>#adding fully connected layer</blockquote><blockquote>model.add(Flatten())</blockquote><blockquote>model.add(Dense(100,activation=’relu’))</blockquote><blockquote>#adding output layer</blockquote><blockquote>model.add(Dense(10,activation=’softmax’))</blockquote><blockquote>#compiling the model</blockquote><blockquote>model.compile(loss=’sparse_categorical_crossentropy’,optimizer=’adam’,metrics=[‘accuracy’])</blockquote><blockquote>#fitting the model</blockquote><blockquote>model.fit(X_train,y_train,epochs=10)</blockquote><h3><strong>The output:</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/922/1*OCIT73UsgFBXP5Uf1A4ejg.png" /></figure><h3><strong>Evaluating the model:</strong></h3><blockquote>#evaluting the model</blockquote><blockquote>model.evaluate(X_test,y_test)</blockquote><h3><strong>Conclusion</strong></h3><p>This article discusses the basic concept of Convolutional Neural networks and their implementation with python code. Since this article was written mainly for beginners, we left more profound concepts like padding, further detail of strides, data augmentation, and other more technical issues not to feel bored. The following articles will dive deeper into deep learning since deep learning is a more profound topic.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aaa315996879" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Air and humidity detection system over IoT using Raspberry PI]]></title>
            <link>https://medium.com/@noman.mazher/air-and-humidity-detection-system-over-iot-using-raspberry-pi-b9e8e020bfed?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/b9e8e020bfed</guid>
            <category><![CDATA[raspberry-pi]]></category>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Thu, 03 Feb 2022 04:35:38 GMT</pubDate>
            <atom:updated>2022-02-03T04:35:38.269Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Air and humidity detection system over IoT using Raspberry PI</strong></p><p>Raspberry pi is the most popular microprocessor used for versatile electronic projects. Working in different sensors like temperature sensors, sea level detector sensors, or humidity detector sensors is Raspberry pi’s power. This article will lead readers to develop Air and humidity detection projects using Raspberry Pi.</p><p><strong>Hardware used for this project:</strong></p><p>Here is the list of sensors and other hardware devices used for this project.</p><p><strong>1.</strong> <strong>DTH11 temperature sensor:</strong></p><p>DTH11, as an essential part of this project, is a very well known temperature and humidity level sensor. DTH11 can be easily interfaced using Raspberri pi. DTH11 contains two sensors in a single pack</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/201/1*IJvgZaOFMbIjUiv0sH-qcw.png" /><figcaption>DTH11 s</figcaption></figure><p><strong>2.</strong> <strong>Raspberry Pi Model 3B:</strong></p><p>3B is the most used model of the Raspberry family. Raspberry pi model 3B has built-in Wi-Fi and Bluetooth connectivity. We will use this to send temperature and humidity information to the cloud server in this project.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1awMHVGZ0t-XH69AKfTheA.png" /><figcaption>Raspberry PI</figcaption></figure><p><strong>3.</strong> <strong>GPRS module:</strong></p><p>Global Packet Radio Service(GRPS) communicates between GSM/GPRS system to a computer. It requires SIM(Subscription Identity Module) as a communication identity.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/210/1*9W0h5qPovQz57j2xOF9XWQ.png" /><figcaption>GPRS</figcaption></figure><p><strong>System hardware design:</strong></p><p>Here we are going to present the hardware configuration of the system. As shown in the figures below Raspberry pi module will work as the centre point of the system. It will get temperature, smoke (pollution) and helpful information from DHT11, MQ135 and LM393 sensors, respectively, and pass this information to the cloud storage with the connectivity of the GPRS module.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/399/1*m_HwmhcWwClAuzSGQR2R-w.png" /><figcaption>Block diagram of system</figcaption></figure><p>To reduce the overhead of data on cloud storage, it will first decide pollution level, <strong>Air Quality Indexes(AQI)</strong> crosses threshold level be sent to cloud storage. The following formula will measure AQI.</p><p><strong>Software setup:</strong></p><p>Raspberry pi can be used with multiple languages like C, java, python. In this project, we will use python language for programming.</p><p>We will present separate code for each sensor for ease of users</p><blockquote>LM393 sound sensor code:</blockquote><blockquote>import LiquidCrystal_I2C from time</blockquote><blockquote>import sleep</blockquote><blockquote>import RPi.GPIO as gp</blockquote><blockquote>gp.setmode(gp.BOARD)</blockquote><blockquote>gp.setup(7,gp.IN)</blockquote><blockquote>lcd=LiquidCrystal_I2C.lcd()</blockquote><blockquote>lcd.clear()</blockquote><blockquote>lcd.display(“Initiating Mic…”,1,0)</blockquote><blockquote>sleep(1)</blockquote><blockquote>lcd.display(“Listening”,2,4)</blockquote><blockquote>sleep(1)</blockquote><blockquote>while True:</blockquote><blockquote>try:</blockquote><blockquote>while(gp.input(7)==1):</blockquote><blockquote>print(“SPEAKING…”)</blockquote><blockquote>lcd.display(“SPEAKING”,1,1)</blockquote><blockquote>else: lcd.clear()</blockquote><blockquote>except KeyboardInterrupt:</blockquote><blockquote>gp.cleanup()</blockquote><blockquote>break</blockquote><p><strong>MQ135 gas sensor code:</strong></p><blockquote>import sys</blockquote><blockquote>import math</blockquote><blockquote>import operator</blockquote><blockquote># The load resistance on the board</blockquote><blockquote>RLOAD = 10.0</blockquote><blockquote># Calibration resistance at atmospheric CO2 level</blockquote><blockquote>RZERO = 76.63</blockquote><blockquote># Parameters for calculating ppm of CO2 from sensor resistance</blockquote><blockquote>PARA = 116.6020682</blockquote><blockquote>PARB = 2.769034857</blockquote><blockquote># Parameters to model temperature and humidity dependence</blockquote><blockquote>CORA = 0.00035</blockquote><blockquote>CORB = 0.02718</blockquote><blockquote>CORC = 1.39538</blockquote><blockquote>CORD = 0.0018</blockquote><blockquote>CORE = -0.003333333</blockquote><blockquote>CORF = -0.001923077</blockquote><blockquote>CORG = 1.130128205</blockquote><blockquote># Atmospheric CO2 level for calibration purposes</blockquote><blockquote>ATMOCO2 = 397.13</blockquote><blockquote>@brief Get the correction factor to correct for temperature and humidity</blockquote><blockquote>@param[in] t The ambient air temperature</blockquote><blockquote>@param[in] h The relative humidity</blockquote><blockquote>@return The calculated correction factor</blockquote><blockquote>def getCorrectionFactor(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG):</blockquote><blockquote># Linearization of the temperature dependency curve under and above 20 degree C</blockquote><blockquote># below 20degC: fact = a * t * t — b * t — (h — 33) * d</blockquote><blockquote># above 20degC: fact = a * t + b * h + c</blockquote><blockquote># this assumes a linear dependency on humidity</blockquote><blockquote>if t &lt; 20:</blockquote><blockquote>return CORA * t * t — CORB * t + CORC — (h-33.)*CORD</blockquote><blockquote>else:</blockquote><blockquote>return CORE * t + CORF * h + CORG</blockquote><blockquote>“””</blockquote><blockquote>@brief Get the resistance of the sensor, i.e. the measurement value</blockquote><blockquote>@return The sensor resistance in kOhm</blockquote><blockquote>“””</blockquote><blockquote>def getResistance(value_pin,RLOAD):</blockquote><blockquote>return ((1023./value_pin) — 1.)*RLOAD</blockquote><blockquote>“””</blockquote><blockquote>@brief Get the resistance of the sensor, i.e. the measurement value corrected</blockquote><blockquote>for temp/hum</blockquote><blockquote>@param[in] t The ambient air temperature</blockquote><blockquote>@param[in] h The relative humidity</blockquote><blockquote>@return The corrected sensor resistance kOhm</blockquote><blockquote>“””</blockquote><blockquote>def getCorrectedResistance(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD):</blockquote><blockquote>return getResistance(value_pin,RLOAD<strong>) / </strong>getCorrectionFactor(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG)</blockquote><blockquote>“””</blockquote><blockquote>@brief Get the ppm of CO2 sensed (assuming only CO2 in the air)</blockquote><blockquote>@return The ppm of CO2 in the air</blockquote><blockquote>“””</blockquote><blockquote>def getPPM(PARA,RZERO,PARB,value_pin,RLOAD):</blockquote><blockquote>return PARA * math.pow((getResistance(value_pin,RLOAD)/RZERO), -PARB)</blockquote><blockquote>“””</blockquote><blockquote>@brief Get the ppm of CO2 sensed (assuming only CO2 in the air), corrected</blockquote><blockquote>for temp/hum</blockquote><blockquote>@param[in] t The ambient air temperature</blockquote><blockquote>@param[in] h The relative humidity</blockquote><blockquote>@return The ppm of CO2 in the air</blockquote><blockquote>“””</blockquote><blockquote>def getCorrectedPPM(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD,PARA,RZERO,PARB):</blockquote><blockquote>return PARA * math.pow((getCorrectedResistance(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD)/RZERO), -PARB)</blockquote><blockquote>“””</blockquote><blockquote>@brief Get the resistance RZero of the sensor for calibration purposes</blockquote><blockquote>@return The sensor resistance RZero in kOhm</blockquote><blockquote>“””</blockquote><blockquote>def getRZero(value_pin,RLOAD,ATMOCO2,PARA,PARB):</blockquote><blockquote>return getResistance(value_pin,RLOAD) * math.pow((ATMOCO2/PARA), (1./PARB))</blockquote><blockquote>“””</blockquote><blockquote>@brief Get the corrected resistance RZero of the sensor for calibration</blockquote><blockquote>purposes</blockquote><blockquote>@param[in] t The ambient air temperature</blockquote><blockquote>@param[in] h The relative humidity</blockquote><blockquote>@return The corrected sensor resistance RZero in kOhm</blockquote><blockquote>“””</blockquote><blockquote>def getCorrectedRZero(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD,ATMOCO2,PARA,PARB):</blockquote><blockquote>return getCorrectedResistance(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD) * math.pow((ATMOCO2/PARA), (1./PARB))</blockquote><blockquote>“””</blockquote><blockquote># Arduino: (0 a 1023)</blockquote><blockquote># Raspberry Pi: (0 a 26690)</blockquote><blockquote>“””</blockquote><blockquote>def map(x,in_min,in_max,out_min,out_max):</blockquote><blockquote>return (x — in_min) * (out_max — out_min) / (in_max — in_min) + out_min</blockquote><blockquote>def main():</blockquote><blockquote>value_ads = 3300 # value obtained by ADS1115</blockquote><blockquote>value_pin = map((value_ads — 565), 0, 26690, 0, 1023) # 565 / 535 fix value</blockquote><blockquote>rzero = getRZero(value_pin,RLOAD,ATMOCO2,PARA,PARB)</blockquote><blockquote>correctedRZero = getCorrectedRZero(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD,ATMOCO2,PARA,PARB)</blockquote><blockquote>resistance = getResistance(value_pin,RLOAD)</blockquote><blockquote>ppm = getPPM(PARA,RZERO,PARB,value_pin,RLOAD)</blockquote><blockquote>correctedPPM = getCorrectedPPM(t,h,CORA,CORB,CORC,CORD,CORE,CORF,CORG,value_pin,RLOAD,PARA,RZERO,PARB)</blockquote><blockquote>print(“\n MQ135 Gas Sensor:\n”)</blockquote><blockquote>print(“\t MQ135 RZero: %s” % round(rzero))</blockquote><blockquote>print(“\t Corrected RZero: %s” % round(correctedRZero))</blockquote><blockquote>print(“\t Resistance: %s” % round(resistance))</blockquote><blockquote>print(“\t PPM: %s” % round(ppm))</blockquote><blockquote>print(“\t Corrected PPM: %s ppm” % round(correctedPPM))</blockquote><blockquote>if __name__ == “__main__”:</blockquote><blockquote>main()</blockquote><blockquote><strong>DTH11 temperature sensor code:</strong></blockquote><blockquote>import time</blockquote><blockquote>import board</blockquote><blockquote>import adafruit_dht</blockquote><blockquote>import psutil</blockquote><blockquote># We first check if a libgpiod process is running. If yes, we kill it!</blockquote><blockquote>for proc in psutil.process_iter():</blockquote><blockquote>if proc.name() == ‘libgpiod_pulsein’ or proc.name() == ‘libgpiod_pulsei’:</blockquote><blockquote>proc.kill()</blockquote><blockquote>sensor = adafruit_dht.DHT11(board.D23)</blockquote><blockquote>while True:</blockquote><blockquote>try:</blockquote><blockquote>temp = sensor.temperature</blockquote><blockquote>humidity = sensor.humidity</blockquote><blockquote>print(“Temperature: {}*C Humidity: {}% “.format(temp, humidity))</blockquote><blockquote>except RuntimeError as error:</blockquote><blockquote>print(error.args[0])</blockquote><blockquote>time.sleep(2.0)</blockquote><blockquote>continue</blockquote><blockquote>except Exception as error:</blockquote><blockquote>sensor.exit()</blockquote><blockquote>raise error</blockquote><blockquote>time.sleep(2.0)</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b9e8e020bfed" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Leads extractor from google maps:]]></title>
            <link>https://medium.com/@noman.mazher/leads-extractor-from-google-maps-c8f65c614e23?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/c8f65c614e23</guid>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Mon, 17 Jan 2022 04:29:40 GMT</pubDate>
            <atom:updated>2022-01-17T04:29:40.978Z</atom:updated>
            <content:encoded><![CDATA[<ol><li>Go to <a href="http://www.maps.google.com/">www.maps.google.com</a></li></ol><figure><img alt="" src="https://cdn-images-1.medium.com/max/975/1*t303WnUYO1g5kZ-6CcuMzw.png" /></figure><p>2. Choose your desired location; for instance, you choose UAE</p><p>3. Choose nearby, for instance; I choose the restaurant (you will not be limited to the restaurant only. Instead, you will select all types of businesses like beauty parlours, schools, shopping malls etc.)</p><p>4. Choose all businesses one by one and copy their phone number</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/975/1*meCRHBw-NDrvd0Q9uwlUKw.png" /></figure><p>5. Copy that phone number to your contacts.google.com or directly copy it in your cell</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/975/1*zgQda8ZVRD_xmZI4WxzoTw.png" /></figure><p>6. Send flyer of our company product as WhatsApp message to each contact</p><p>Note: you can also work with google chrome extension “leads extractor,” but its free version extract only the top 4 leads while we need contact info of all businesses of that country</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c8f65c614e23" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Shared responsibility model for Microsoft Azure]]></title>
            <link>https://medium.com/@noman.mazher/shared-responsibility-model-for-microsoft-azure-508f24d55a86?source=rss-2936a2f6f4d------2</link>
            <guid isPermaLink="false">https://medium.com/p/508f24d55a86</guid>
            <dc:creator><![CDATA[Noman Mazher]]></dc:creator>
            <pubDate>Sat, 15 Jan 2022 04:59:47 GMT</pubDate>
            <atom:updated>2022-01-15T04:59:47.355Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Shared responsibility model for Microsoft Azure</strong></p><p>A cloud computing provider such as Microsoft Azure has to ensure the accountability of its services. To ensure accountability, a framework provides rules and obligations to be followed by the cloud service provider and the users of said services. Such a framework is followed by Microsoft Azure, called <strong>Shared Responsibility Model</strong>. The shared responsibility model is typically followed by an enterprise that controls and manages its own IT Infrastructure.</p><p><strong>Why does Microsoft Azure follow SRM?</strong></p><p>The shared responsibility model dictates the services provider about rules and regulations which are to be followed to ensure the physical security of the data. By the term ‘shared responsibility, we mean that ensuring accountability is the responsibility of the service provider_ Microsoft Azure, in this case_ and the users also share some responsibility for managing data security.</p><p>The shared responsibility model is not a form of the physical structure per se. Instead, it is followed by some cloud service models such as infrastructure as a service (IaaS), platform as a service (PaaS) and software as a service (SaaS). These cloud service providers inform the service provider about the responsibilities of different personnel in the service cycle. They dictate the users and their service providers about their commitments regarding data’s accountability and physical security.</p><p><strong>How do the cloud service models work?</strong></p><p>Cloud service models vary in their role for a cloud service system. For example, one of the cloud service models, IaaS, dictates the users’ responsibility in managing their operating systems and the applications required to stack the data over Microsoft Azure. The users have to make sure that their systems (mobile phones/ laptops etc.) are working correctly, and there is no anomaly in the system. Users also have to download and install Microsoft Azure to upload and manage their data. These are the responsibilities of cloud users. Now, let’s talk about the responsibilities of the cloud service provider, which is Microsoft. According to Cloud Standards Customer Council (CSCC), the Microsoft Azure service provider has to work with virtual machines and provide said resources to the users. The provider is also responsible for the security of the physical safety of data centres, disks and networks etc. in short, the users are responsible for security in the cloud, i.e., their data. In contrast, the provider ensures the safety of the cloud.</p><p>The responsibilities vary according to each cloud service provider and the cloud service models. Hence, there is no physical form of shared responsibilities model as stated earlier.</p><p>Regardless of the cloud service model and provider, some responsibilities are fixed according to every model. For example, the users/ customers will always be responsible for their information and data (what they upload on the cloud and who do they share it too), devices (mobile phones/laptops/ PCs etc.) and their online identities and authentication information (password and IDs).</p><p>Similarly, the cloud service provider will always be responsible for the physical host (i.e., the virtual machines etc.), network and data centres.</p><p>Some responsibilities differ from cloud service model to model and provider to provider, such as; software stack, directory infrastructure, network controls and operating systems.</p><p>The primary duty of the users is to protect their identities and the data they upload.</p><p><strong>Microsoft Azure Service model:</strong></p><p>Microsoft provides services based on on-premises software products. It is convenient for the users. The providers can have strong access to the services, shifting some of their data security responsibility to the users.</p><p>However, Microsoft Azure is a public model, which means the users only have a web browser to access the services and resources. Microsoft owns all the hardware, software, and infrastructure resources.</p><p>SQL Azure provides services related to data management. It includes a database for the users and facilities such as data synchronization.</p><p>Microsoft Azure typically hosts under the cloud service model PaaS. It provides windows, database and development tools like NAPA. SharePoint online operates under SaaS while virtual machines, networks and storage are provided according to the IaaS. The user can deploy infrastructures like virtual machines using the development tools. With the PaaS model, Azure enables users to run theirs on applications or other third-party applications on the operating system provided by Microsoft. Microsoft Azure also provides a service management environment. There are different roles that Microsoft Azure in PaaS can support, such as worker and web roles. Web role is used to run and control web applications while idle tier applications are run and controlled in a worker role. The users do not have to worry about software updates. The Microsoft Azure host updates all the software regularly. Moreover, the host also upgrades the hardware resources according to the defined responsibilities in PaaS.</p><p>With the SaaS Microsoft Azure, the user can access software from various devices. Only the software that is needed for the user’s business is accessed. The provider manages everything from e-mail and CRM to infrastructure and operating systems. SaaS saves each user as a <em>tenant</em>. Each tenant works with a personalized virtual application. In this way, it makes a multitenant architecture.</p><p><strong>What is Microsoft Azure Stack?</strong></p><p>The Microsoft Azure stack follows an on-premises service model that enables the users to control better their data and access resources that Microsoft provides. The On-premises model extends services and resources to the user’s choice environment.</p><p><strong>Conclusion:</strong></p><p>Microsoft Azure provides services according to all the cloud service models in addition to on-premises products. It provides a great deal of security using shared responsibility. The workload of provider and user is balanced so that each person has to worry about their share of responsibility, and the rest is managed by the other. The users have to take care of their data, and the provider takes care of the resources and protects user credentials.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=508f24d55a86" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>