Getting the information you need should be as easy as asking for it, here’s how to do that using our app.
Here at Chata, we believe that databases, no matter how large or complex, should be easy to access for anyone looking to gather information and glean helpful insights from their data. That’s where conversational data exploration comes in. Instead of navigating through spreadsheets or drop-down menus in cloud software systems, we’ve made it possible to simply ask for what you need and have that data served back to you, immediately.
Imagine this: you’re at your favorite restaurant and after browsing the menu for a couple of minutes, you decide to order some pasta. The waiter comes over to your table and you tell them you’d “Love an order of the fettuccini alfredo, please.” The waiter puts your order in and a few sips of your cocktail later, the pasta arrives.
In the real world, you’re used to asking for what you need from other humans and receiving what you’ve requested. When it comes to interacting with your computer, however, making requests for what you want looks pretty different. We’ve all been trained to navigate drop-down menus, hit Enter, navigate through different windows or tabs, and click on icons to get things done.
We developed the Advisory Studio and Performance Studio by Chata to close the gap between wanting information from your data and actually getting it, the same way a waiter closes the gap between you wanting a bowl of pasta and you actually receiving it. Without the waiter (and the chef behind the cooktop), you’d have to go digging around in the kitchen for ingredients, put the water on, and start whisking it all together yourself.
Using Chata, getting answers from your data is as easy as asking for what you want. Conversational data exploration is the process of accessing a database using the everyday language you’d employ in a conversation with another human. This is the core foundation of our app, and of every feature that lies within.
But, we know that when it comes to interacting with data, asking questions of a computer isn’t necessarily second nature for most of us. In our series, How to Ask Effective Questions in Chata, we’ll cover some strategies for getting into this mindset, and we’ll share tips about how to ask great questions to help you find exactly what you’re looking for.
Be Clear, Be Specific
In Chata, data exploration begins with queries, which are questions or requests that pertain to your data needs.
The clearer you are when querying your data, the easier it is for Chata to interpret your query and return the exact information you’re interested in. Returning to our earlier analogy, Chata’s system operates much like a waiter for your data. You ask our app for information, it searches through your database to collect the data pertaining to your question, then it retrieves and returns what you requested, all within seconds.
Let’s take a look at an example:
Say you’re interested in knowing more about your sales data. You could pop into Data Chat and type in “sales.” What you’ll get is a number that reflects all sales: a list of every single sale stored in the database. But that might not be particularly helpful if this comprehensive list isn’t really the information you were after. You need the data that gets returned to be useful and actionable, and a wide-ranging summary isn’t likely to yield deep insights or help you understand the nature of sales under specific circumstances.
So, you need to consider exactly what you’re interested in knowing. Your query should reflect that.
Returning to the restaurant analogy, if the menu listed 10 types of pasta, you would need to specify which pasta you wanted when your waiter took your order. Alternatively, you could say “I’d like pasta with cheese on it” and the waiter might help you narrow your search down to three kinds of pasta that meet that criteria. Then you might say, “I’d like a pasta with a cream sauce” and the waiter would then be able put you down for a bowl of fettuccini alfredo.
In Chata, the more specific your query, the more specific (and useful) the answer the system can provide. Because it’s a computer, it can’t turn around and ask you “Which sales information do you want, exactly?” But, by entering a clear, context-rich query such as, “Total sales by month this year,” or “All sales by rep this month,” or even something like, “Total sales by month by rep this year,” Chata is able to return the exact information you’re really after, within seconds.
If you come to the system with your question narrowed down in this way, you’ll see the results you’re expecting and you’ll be equipped to explore further or make decisions based on that information.
Again, it’s not so different than asking for information from another human. You probably perform conversational data exploration with other humans on a regular basis! For example, if a CEO was looking to find out more information about the recent performance of the sales team at his company, he might approach the Sales Director to find answers. If the CEO went up to his Sales Director and simply said, “Sales,” what would the outcome be? It’s likely that the Sales Director would be confused (or even irritated at the vagueness of the statement) and respond with something like, “What about sales?” or “What information about sales are you looking for?” To have an effective conversational data experience, context is absolutely critical.
When querying your data with Chata, it’s helpful to think “If I approached another person who had knowledge about this topic and asked them to answer this question, would they be able to answer me without further clarification?”
Leverage the Value of Value Labels
The second level of specificity is ensuring that you’re using the correct value labels during your data exploration process. Value labels are essentially used to name a group of data within a database, so it can be categorized or organized properly. In the pasta example, value labels ensure that there’s clear communication about “spaghetti carbonara” versus “spaghetti Bolognese”. To make sure that Chata is returning data about the correct “spaghetti” (that is, any data point or group of data points in your database), you need to provide an accurate value label.
Sometimes, value labels don’t match with exactly how humans refer to that thing in real life. For example, a restaurant may be called “Maria’s Kitchen”, and this is how most everyone refers to it in day-to-day conversations. But in the database, that company might be labelled as “Maria’s Kitchen Restaurant Ltd.” This is the value label associated with the restaurant in the database.
If you’re unsure what value labels you can query about, it’s actually very easy to determine which value labels are being used in your database. In Chata, simply query “All [value label variable]” and you’ll see a full list of the words the computer knows to recognize. This includes dimensions like “All customers”, “All inventory”, or “All locations”. Databases sometimes contain multiple value labels that share similarities, so it’s important to use exact value labels to help the application know exactly what you are querying about.
With that said, our system is also capable of inferring context around value labels. If a customer and a vendor share similar value labels (e.g. customer: Maria, vendor: Maria’s Kitchen Restaurant Ltd.) it will interpret your query and return data relevant to either the former or the latter, depending on the context you provided. Again, greater specificity lends itself to better results.
Additionally, if you happen to misspell a word or enter an abridged version of a value label (as above with “Maria’s Kitchen” versus “Maria’s Kitchen Restaurant Ltd.”) our system will catch similarities and verify what you intended to enter. To check that your query was understood correctly by Chata, you can always click the little robot icon beside the answer that was returned to you to see how the query was interpreted.
These are just a handful of features we’ve built to help support the human-computer conversation. Just like the Sales Director asking for clarification and context from his CEO, Chata is built to ensure that it’s accurately returning the information you are looking for.
In Part 2 we’ll go over how to create more complex queries that get you straight to the detailed answers you need to make better data-driven decisions, faster. Read more in our next post, How to Create Queries for Chata →