Visualising The Profile Part 3 Watson Personality Insights

Key Takeaways

all righty what's happening guys welcome to part three in the series on how to use IBM personality insights to create better customer profiles so in the last video of what we did is we created a customer profile for none other than mr. Warren Buffett and we looked at the different parts of the profile that are returned by the service so what we got back are the Big Five personality traits and what percentiles that he ranks against those we looked at needs and we also looked at values so out of that you actually get quite a fair bit of insight into that particular person's personality now what we're gonna do in this video rather than just sticking and analyzing a JSON string we're actually going to throw it all into a data frame and we're going to visualize it using Seabourn and matplotlib so without further ado let's get right into it alrighty cool so jumping back into our Juvonen notebook so we've already got our profile here so if we take a look at our profile again you can see that we've got all of that good stuff and let's just check so yeah all right cool so we're also good so got it all in it now what we're going to do is we're going to start playing around with Seaborn so I guess to get a little bit better visualization going on here let's just create a section and start segmenting this off so we're going to call it visualizing profile and alrighty let's start importing the stuff that we need so we need to import pandas let's import pandas as P do you and then we're going to also need or are we gonna need we're gonna name up four babe to work with Seaborn so important that but yeah babe import pie plot as PLT sorry I should have added Seaborn into there as well Seabourn imports OC one just makes visualization just a little bit nicer right I'm going to import that there's SNS alright cool so that gives us the three key libraries that were sort of going to need for this next step now what we're gonna do is we're going to just visualize the needs bit for now so let's rip that out throw it in a data frame so that Seabourn can accept it and let's start visualizing so if you remember our needs feature is stored inside needs and you can see that here it's just stored in an array so what we can actually do is extract that out and turn it into a data frame so what we'll do is we'll actually go we're going to use a comprehend of this so let's for need in profile that needs and we are going to return back a another array so we're gonna return a dictionary so we want need and we really just want to extract the need name so we want here we'd be wanting challenge and we'd want the percentile as well just to see what what we actually or just to get back what we exactly want so we're going to and we're going to associate the name with the percentile so what we'll do is we'll create a need name so that would be our key and then we're going to use or assign a percentile to that fit alrighty so this should return need name need a percentile and let's just check with formatted that correctly sonido name need a center that should be fine alrighty that looks good and let's just store that in a result variable let's take a look at that perfect all right so we've got a dictionary perfect that has all of our need results good now what we're going to do is we're going to convert that into a data frame so let's call data frame PD data frame and we're going to use the from date method to load this up so we're going to pass through a result and we want to create an index initially so we're going to say orient equals index okay so we've got our data frame and what we want to do is just rename our columns that's right let's reset the index because we want to rename our columns and we're going to say in place it goes true perfect and now we've got no indexes of what will then do is won't brain Emma columns and we're going to call it a need and the center perfect all right so we've got app column down now what we can do is we can actually visualize this with Seaborn so let's do that so let's create a bar plot and we're going say y equals n equals at a percentile and x-axis equals made and the data equals the data frame so and then what we're going to do is run that show perfect are not so perfect looks like it's a little bit small so let's change this figure size and we'll say figure size equals 1 why does so let's try that all righty cool so you can see immediately that's a much better way to visualize the actual need or the the trade profiles or God clearly a high degree of structure practicality curiosity I mean - yeah that I'm probably associated that with warm baths Challenge and practicality is a very practical invested value-based but you can instantly see that you're actually getting some really interesting insights by by doing this type of thing now what we can do is we can again visualize some of the other traits but what we're actually going to do in the next video is we're going to just create a quick function to loop through all of the traits or so that we can pass a particular trait and it it'll visualize it for us simply automatically alrighty guys that about wraps up this video if you found it useful be sure to like share and subscribe you're doing me a huge favor peace