It's easy to load data in to MongoDB. In part, this is because MongoDB supports a very flexible data model. Meaning, you do not need to define a schema in advance or set up any tables. Tables, by the way are called collections in MongoDB. In addition, MongoDB is flexible about the shape of individual documents within a collection. The number of fields, specific names of keys, and value types can vary from one document to another within the same collection. In general, you should develop and use a consistent data model but MongoDB makes it easy. To use an exploratory iterative approach to developing the data model. This is important in nearly all use cases for data science. Requirements evolve and the availability of new data sources imposes changes on our data models. So let's get started with an example of loading, cleaning, and iterating on a data model using MongoDB. First, we're going to load some data into your free tier cluster, for this we're going to use the Mongo import program and import the csv file of the data for a little more than 46,000 movies. First, download the file movies_initial.csv from the handouts, then we will import the data, using the mongoimport command. Mongoimport is distributed with MongoDB. Since you have already downloaded and installed MongoDB, you have access to it from the command line interface of your choice. On a Mac or Linux machine, you'll probably use a Bash shell of some kind. On Windows, you'll probably use the Command Prompt, or PowerShell. For this example, we're going to use some movies data from the open movie database also known as OMDB. This data set is nice because it is large enough and requires enough data cleaning to be realistic. In addition, there are a number of other publicly available data sets that we can work to integrate with the OMDB data. In a manner that effectively simulates the types of challenges we face in developing and maintaining data resources to support enterprise analytic applications. To import our initial data set you'll use mongoimport, specifying several command line options. Here's the template for the command you'll use. You simply need to fill in the command with the cluster name and seed list for your free tier cluster. To get this string, go into Atlas and from the cluster view, click the CONNECT button. In the model that pops up, select Connect with the Mongo Shell. Mongoimport has a slightly different API than the Mongo shell. So, we just need to copy a couple portions of this connection URI. First, we'll copy the cluster name. The cluster name for your replica set should appear as the value for the URI parameter replica set. Go ahead and copy and paste it into our mongoimport command. Replacing the cluster place holder with the actual cluster name. Next, we need to select the seed list from this URI. The seed list begins immediately after the :// in this URI and ends just before /test. The seed list is nothing more than a list of the fully qualified domain names, plus the port on which MongoDB is running, for all the servers in your MongoDB cluster. Replace the seed list placeholder with the seed list you've just copied. See the lecture notes for this lesson for a mongoimport template command you can copy and paste. And for a written version of these instructions. Once you have correctly added your seed list to the mongoimport command go ahead and run it. But before you do, make sure that the movies_initial.csv file is located in the same directory from which you're running the mongoimport command. I can run exactly the same command, in Windows, from the Command Prompt or the Windows Shell. Mongoimport has the same API, regardless of which operating system you're running on. Once mongoimport finishes running, you should see a message indicating how many documents were imported.
