Data Analysis with Google Earth Engine | Alemayehu Midekisa | Skillshare

Data Analysis with Google Earth Engine

Alemayehu Midekisa, Geospatial Data Scientist

Play Speed
  • 0.5x
  • 1x (Normal)
  • 1.25x
  • 1.5x
  • 2x
4 Lessons (31m)
    • 1. Earth Engine Up and Running

      5:52
    • 2. Coding on Earth Engine

      9:26
    • 3. Code Editor

      6:06
    • 4. Lab 1: Earth Engine

      9:31

About This Class

This course provides both theoretical knowledge and practical skills in Advanced Satellite Data Analysis with Google Earth Engine. In this course, you will be given hands on practical exercises to master analyzing big geospatial data on the cloud. You will learn to access, process and analyze satellite data including Landsat, MODIS, and Sentinel and others using an open source platform. You will also learn to classify satellite images using machine learning algorithms. You will also have access to the lab exercise scripts as part of this course. 

Transcripts

1. Earth Engine Up and Running: Hello. I'm a doctor, Alama. And welcome to big Geospatial Data analysis was Googlers engine in this course, I'll teach you how to visualize processing an ally Saleh Data using a freely available clock platform tools. Let's get started. So the first thing first ours engine app and running. So if here, I'll show you how to sign up for Googlers Engine and get started was coolers Engine So first Googlers engine is a platform for petabytes scale scientific data Analysis on visualization A swell for boasts probably come and, um nonprofit a nonprofit Willis Government users and others injured store satellite imagery with organizes and makes it available for the first time to be able to do planetary skilled data. Analysis is on the cloud Eso The Public Data archive includes historical ours imagery going back more than 40 years and also new miseries updated on the coolers. Engineer Kaif us. It is available. Anders Engine also provides an A p I application programming interface in the JavaScript as well a Spuyten And there are also other tools to enable the an ounce off large geospatial data. So another interesting feature on Googlers engineers A time lapse which is a pretty cool application. What it does is it provides you a time serious are ah change map for any given place on the planet showing using Lancelot Satellite Data Archive. So technically, what is is showing is animal change over time, starting from 1984 to 2016 for any given place on Earth. So to show you these, I'll have some examples so that you can visualize. Oh, are you can see how the's time lobs. So the first example is this Copy this link that I prepared for you and open a new crime brother New crawling window. And then here's what you have. So this is a time serious off change on on Dubai. As you can see here from it starts from 9 1984 goes to 2016 so and will change over time. What happened since 1984 in Dubai, you can see changed from a totally desert area to a mega CD. A big infrastructure development e over over those parents. So which is pretty interesting to visualize thes change over time using launched a satellite data. Another example that I prepared today is here. I'll copy that and pissed and then open a new window using chrome. So here is a San Francisco Bay area in California showing change over time using Lance, a satellite data chickens here from 1984 to 2016. You can see here in Auckland area how special patterns changing over time, new construction and development. Um, in San Francisco. Swell. Which is pretty interesting to monitor that change over time at the quick off a button using this large for him off a satellite data. The beauty of that is you can see these change over time for any given place owners. This is another third example that I prepared today. This is Las Vegas. As you can see here from 1984 to 2016 there is a big expansion in in Las Vegas. You can see you know, the satellite data captures that change on infrastructure expansion in the city of Las Vegas, in Nevada, In USA, the last example that I have for you is let me just open there. So here's an as our example. Um, so this is in the amount of basing in Brazil to show some changing forest covers. You can see here from 1984 to 2016. There's some changing forest cover that means deforestation, clearing of forest here. So the application off this change over time. Vars You know, it's very important to use it in different applications ranging from first re, you know, urban infrastructure on D's is and what have you so pretty cool application? 2. Coding on Earth Engine: some when he opened the Googlers engine plot firm. This is what you see on the main Bage planetary scale plot firm for Earth science data and analysis. Powered by Google's cloud infrastructure, you can watch a video here in some additional information about the Googlers on. So here are the Mainers Burglars Whips Man website. Here's frequently asked question time lapse while which I already I already explained earlier data sets, which I'll discuss kids steady, some interesting application that used coolers, engine and plot firm and a sign up for to sign up for Googlers engine. So let's see. Well, let's go to the coolers in general upside and see what's what's up there. So to do that, I'll open a new Kirm window and ty good go spurs. And here is the Googlers Engine Man website, which, as time lapse Googlers, Injun Explorer data sets plots firm in case day. So this is the exact website earth engine dot google dot com. Okay, I'll click that here is exactly what I show earlier. So here you can scroll down and see some, you know, information. More information about Googlers engine the Googlers engine cloud computing platform so most importantly here frequently asked question. You can read that for your information and the time lapse is this'll. So in this time lapse, you can search for any place on earth on the planet, and then you can. Once you have that location, you can see visually the change over time from 1984 to 2016. So, for example, this is Brisbane, Australia. Click that there. I'm some already said app. Places for the time. Lob soon solved. Just click one off this, which is Brisbane, Australia, and which sounds to be a little bit slower. So once I have that, what I can do is I can a share in bits. I click that, so it gives me to share a link or in bed a time lapse in your website. So if I click this and I can share this link to other people, email that or to share that on Facebook or just open it myself in a new browser which should bring the same location so you can share the cool thing about time lobsters. You can share in bed that in your website or shared the link to other people. You know friends and colleagues to share this changing dynamics on any place on Earth. So it's a pretty cool application. And Dennis, it's so if I click the data sets, it will give me a least off Davis's that Earth engine archives. It includes imagery, geospatial, data, climate and weather data and democracy. Let's see the imagery. So we have lots of data The Sentinel data motives, moderate resolution imaging, spectral Radiometer, the some high resolution satellite data and in nighttime light imagery data. If you go to the Geospatial data, you can see terrain data digital elevation models. Data certainly made her a star Tim Data of the Global Scale. To open that, you can click thes sir train data and there's Engine. And there's also land cover data in a crop. Plant data used an S Corp plan data lier lan surface temperature data said, and also, um, other geophysical deficits, for example, of global Forest Watch forced data and 1/3 1 is climate and weather data, for example, here we have the most freak data and whether did and some climate data, for example, the agreement and inhale dust, too. And the last one is a demographic data, which includes a world pope, which is a population density mortal estimate off population for the entire worded. And also here Malaria data by the Oxford Miller Last project works for the university. Um, you can you can also access that. So we have Ah, pretty big least off data sets, including imagery, Geophysical climate on whether and demography. Uh, you know, human settlement data. And in some, if you click this main, uh, you have kids. That is some interesting case studies that used ours. Engine application. Finally, here you have overview the coordinator most importantly for the A P I. If you click the court Iager, here's your coordinator. You can right here a JavaScript programming script. And here, once you have that script, you can visualize the map result. And on the right here on the console, you can see some outputs in feedbacks off your your script. And the explorer is literally a graphic is during for interface version off urge engine. It's not not like, according to AP, I you programmatically access data and analyze data. Here. You have data you can access data just ah g y. And the last one is documentation you can access some information if you need to read furs . Or, um so the next one is blawg. You can you can access that. Um you can access that. Here's some interesting blocks. Um, and also here, um, sign up. If you want to sign up for a Googlers engine, you click vis and you'll go toe a sign up for for hours engine. And yeah, so the website is sign updated er Zengin dot google dot com. So, once you you sign up, you feel this firm, um, Google will contact you in the next two days to get access to the Googlers engine plot for clock platform, cloud computing plus firm. Rather, you have to feeling this firm using this website Sign up voters engine dot google dot com and Google. They're not accepting all applications, so you have to feel feel out older feeds as business. You can't says that they really value your request. So mostly Google provides free access to these cloud plot firm the coolers engine platform for research purpose for non profit public use. So you have to justify that you have some project or research that you're going to do and then you need that access to analyze and visualize your geospatial data. So based on that request, they most likely may give you access. Eso. If you're accepted, you'll receive an email from Google and titled Google Groups. You're being added to Googlers. Engine invites within most lately you know, two days, 48 hours off applying. And if you have that email, police read that email carefully because it's it contains further steps how you'll Prasit on Access Googlers Engine plot firm. 3. Code Editor: hello. And this lecture will get started calling honors engine using JavaScript language. So before we start, I would like you to go to this website. I assume that you already have Googlers and an account by now, so you'll start cording using JavaScript. And to do that, you need to log into this website called dot er Zengin dot google dot com. So this will take you to the code editor platform where you will write javascript program to run your analysis on our Zengin. So if you have, if you already open the quote a teacher using the previous script thesis what you're quoted eater would look like. So you have a couple of things here you have to know. So this is the bottom. You have the map canvas will your once your any of your analysis, you'll see your mouth results here at the bottom at the top. Here, this is ah, script accorded eater script platform where you write your javascript language. And here on the left, there are a few things script where you can save your script. Us a private or US A. Shared on an annual several script for future use and then some documentation, if you will. On the right side, you have counsel wants your write a script and run it here, and you can see some information related to the mated it information about your image or your object and some progress off your your script. You can see some. You can monitor feedback here, and there's also a couple of other things on the council here. And Inspector, for example, if you have a mop here, if you click on that specific point, you'll have a value for those points. Ah, here on the Inspector and the task this button years to see what's going on. If you have currently running analysis on Arvin Jinich and monitor, if they're complete or in progress and things like that, we'll see all that in the upcoming lectures. But just, you know, I just want to give you a heads up. There are a couple of things over here. Um, you can once you write a script here, you can hit a run a button to run your analysis, and it can also safe here. You can save your script here for future use, and I lost Think here, if you want to share this a script that you have already written. You can share it with other people by generating get link. So what this will do is would generate a link says that you can share it with other people . When other people open that link, they will have to same script that you have so kind of improving collaboration among, uh, other people in France and colleagues. So so hyla worthy JavaScript. So a standard You know, when you will start by writing a pre intel awarded on a semi colon, JavaScript needs a semicolon at the partisans in a quotation off these hell worded. So let's see. Let's see if we write up this in and print that inner Zengin so kind of a test rent. So what I'll do is I'll go to the code a teacher. I will open any window right court, since I already opened earlier. Have a safety, you know, link here. So I'll just click called daughters engine google dot com so you can see here called daughters engine dot google dot com. So if you click here, you can you will have a new window called a teacher windows. I mentioned earlier. Here's a script, documentation and things like that. And here's Accorded Eater, where you're writing a script and a council in the mob canvas. So let's right, let's start writing our first JavaScript long gauge. So before I do that, I always have a habit off writing on a comment. What? The script. Death. So this is technically print. It just prints Prince Hell awarded. So I'll have to do this in a small cap because JavaScript is sensitive to capitalization. So print and then apprentices. So just have a coated opening quotation in right Hello Were. And if I just close the partisan and Sammy Colin So this is my first brand new JavaScript long gauge. I'm gonna test run it honors engine eso. What I'll do next is once I write my script, I'll just hit Run here and then when a heat run, it should show me some result here on the console. Let's see what's going on. I just hit. Run! You see, I have my print Hello word It is printing here on the console that successfully doing that great. And if you do the same thing just right print Taylor worded in your court a teacher on. Then he run. You'll have the same result 4. Lab 1: Earth Engine: Hello? Uh, this is LA one exercise for U S. Oh, well, diving into Google Earth engine ap I programming using JavaScript on language so we'll dive into do some quoting. So once you start quoting honors engine, um, you'll have to copied if the following link for the script. Or you have to click the link below for Lab one, which is attached below the lecture. So once you run that script on Googlers engine AP, I you'll have a new image that looks like exactly these. So what it is is the A collection, off land SAT satellite image processed for the entire United States, and it's a true color compose. I'd combination off different bands off the satellite sensor and mosaic over time and space . So you have something like this If you, uh, exactly do what you know and and run the script and do us it is distracted. You'll have something like like this pretty cool map, so it's good to exercise. Lob Juan quote. So this is my, um, quoted teacher plot firm on the left side, as you can see here, have some scripts already saved, which I don't need. Now. I don't need for now. And here is the consul Where the the script causal, where you can write your script, a JavaScript programming language. And on the right, there are three things, Inspector, Consul, in tasks. So mostly the console you would. I need these to monitor and see some feedback or just get information about your data, meta data, information and things like that. And most importantly, the map canvas on the bottom here would provide you some map result to visualize once you run your analysis honors engine. So once you run your called, you'll have your output map here you can visualize, and then you can check in and check here. Sil, before we ran the script, let me give you, Ah, a little bit off a demo description about the court. So first thing, first you like, I would like to provide some information about what the court is in this case. Love want lands a satellite image for USA, and then who? The altar is some contact information. And you know you can do these or you can opt out required so that my first line of court is this'll one. So what it does is it tries to, um import fusion table, which has countries boundary for all countries off the world. And so I'm defining variable cold countries here And because the feature collection I'm calling a future collision and providing that FBI d so that what that death is, it will bring import that data countrys pondering data. And the second line is I'm defining a viable called us A. And what I'm ordering is just filter a country called United States from this country's boundary. So I'm filtering Ah, certain country in this gives us a And the next line, of course, is importing the landside eight collection from Googlers Engine Data Archive. So I'm defining a viable coal collection here. And I'm calling the image collection specifying the land sat. And because there are different lines how centers land Landzaat sensors Lance at five, seven and eight. In this case, I'm calling Elsie eight, which is Lance at eight and then specifying the right image collection data time. And I can also filter bound by us already defined what us is here. Right? So this is a boundary country boundary for USA. So I feel I want to filter the image based on this boundary, I can also define or feel turn by debt using a filter that, ah, function. So the filter that function will provide me to filter my image collection based on dates so that the 1st 1 is a start date for that image regulation. And the 2nd 1 is unending dead. So my did My image collection will start from distant to this debt. That means January 1st 2072 April 30 2017. So you can define any date that you want by modifying this, filtered it and also country dead. And the next thing is because it's a like Lan Sad has stayed off every two weeks. So we have, you know, several image in this data. So we What we need now in this script is to generate kind of reduced that image into a single map by running and minimum stats sticks over the entire image layers. So here I'll compute a minimum statics over that entire image collection and have technically a single map. And the next thing is to display that image in a true color comp aside here. Ah, combination Off bond for Bond three and bunt off Lanza. Eight. Kind of fine, a maximum value. And this is also the title that's us. You'll see here on the layer. So Lan sat sat lite for U. S. A. And then the last one is a map center Object will define the zoom to which you want to your map out to center on. So here, in this case, I'll center it in USA USA is already defined here in the second line and now I'll run my skip to run your script. You just click the run button here. I'll do that. So when I read that, it just gives me, um, a result. So it competes all these things and in a fraction off a second, this large volume off satellite data. It just runs on Google's machines, tens of thousands of missions and Parle computing platform, which, in a traditional desktop competing this should have taken, you know, days literally so tanks to Google for providing this Googlers engine cloud computing platform. Now we're able to process this large volume off satellite data for the entire United States over time and space in computing sem analysis and visualizing result instantaneously almost so, which is very, very cool. So this is our land sat satellite data alongside eight data, which is we mapped it in a true color composite. What it means is that it will provide us ah, visualization or color asi top years in their reworded. For example, water is blue, you know, vegetation is green. As you can see, soil is, you know, brownish on and thinks like that, so pretty straightforward and intuitive to visualize. So you can, you know, walk around any place that you're interested in the US and this map. So for for some reason, if you're interested to lock, it's a place you can check these and and check to add and remove the satellite data from your zoom window, and then you can go to any anywhere You're interested for this example. It's good. You know, San Francisco area Bay Area, for example. Here, this is the map background and for San Francisco Bay Area. Now I can add the lance a satellite data by checking this box, and then it will automatically added, and you can remove it