Data Flows

Data Flows are custom connections. Conceptually, Data Flows work like functions: they take in input, process it, and return the resulting output (if applicable). Data Flows are primarily used to connect the front-end of Airkit apps to back-end systems, be they are part of Airkit (such as AirData) or external (such as Salesforce). Any data processing that must occur in order to make these connections smoothly will also be done within a Data Flow. For instance, if an API request returns a JSON object from which only a single attribute is worth displaying to the app user, it is within a Data Flow that the relevant information will be isolated.

For a deeper discussion on Data Flows, see Data Flows. For more information on the different kinds of Data Operations available, see the Data Operations Overview.

Working with Data

When a value is bound to a variable, that value is tied to an individual Session. This simplifies the process of keeping track of information in the short term. If our contact form becomes published, and both Alice and Bob fill it out at the same time, the value of name_input will be different in Alice's Journey than in Bob's. Airkit will handle the details of keeping them separate.

However, outside of any Session, we, the app builders, might want to access certain information collected from all uses of our application. Upon sending out a Contact Form, for instance, we will likely want to be able to access information submitted by everyone, on every Journey, and thus we want to collect all relevant information in one easily-accessible place: AirData. (Note that while, in the case of creating a simple Contact Form, all bound variables are relevant, this will not always be the case. AirData App Objects are customizable precisely so that builders can store information in the way that is most conducive to their use case, rather than being bogged down by automatically-collected but ultimately cumbersome data.)

For a deeper diving into how data is conceptualized in Airkit as well as the crucial difference between binding data to a variable and saving it to AirData, check out Working with Data.

Actions

An Action Chain is a series of Actions taken sequentially.

Actions are a broad category. They can be do everything from run a Data Flow to navigate to a specific Web Page to create a timer. There's also the Condition Action, which makes it possible to create branching Action Chains, where different Actions are taken depending on whether or not certain conditions are met.

For more on how to created an Action Chain, see documentation on Action Inspector Tab, which is where Action Chains are defined. For more on the different kinds of Actions that can be taken, see the reference documentation

Airkit Events

In the context of the Airkit platform, Events refer to any catalyst that might trigger an Action Chain. All Action Chains are associated with a particular Event. It is when – and only when – an Event is fired that its associated Action Chain is performed.

At its core, an Airkit Event is a piece of information that tells the platform something happened and might include optional metadata about the Actions that will consequently be taken. Not all Events will be associated with an Action Chain. An Event merely signals that the platform must check and see if there are any Actions to take. In this way, an Event like a letter: it's sent from somewhere, but ultimately once it gets to the recipient, the recipient decides what they do with it. The letter might spur some action, or it might be thrown directly in the trash.

When we say that Events tell the platform "something happened," we are being deliberately vague. All kind of things might trigger Events. A user clicking a Button Web Control is an Event, as is the instance before sending an SMS message, as is updating a Web Flow.

There are many different types of Airkit Events, including Starting Events, one of which you've already created in the Journey Builder, and Control Events, including the 'Clicked' Event associated with all Button Web Controls.