Introduction

A wrapper project for curious learning analytics.

Setup

Simply add as dependency.

$ npm i @curiouslearning/analytics


Usage

``ts
import { AnalyticsService, FirebaseStrategy } from '@curiouslearning/analytics';

const analytics = new AnalyticsService();

// each strategy will have its own set of configuration options.

// using firebase strategy
const firebaseStrategy = new FirebaseStrategy({
firebaseOptions: {
apiKey: 'api-key',
authDomain: 'my-domain.firebaseapp.com',
databaseURL: 'https://my-db.firebaseio.com',
projectId: 'project-id',
storageBucket: 'my-bucket.appspot.com',
messagingSenderId: 'sender-id',
appId: 'app-id',
measurementId: 'G-measurement-id',
},
userProperties: {
campaign_id: 'my-campaign',
source: 'my-source'
}
});

await firebaseStrategy.initialize();
analytics.register('firebase', firebaseStrategy);
analytics.track('initialized', { test: 'test' }); // fires a tracking event to firebase

// using statsig strategy
const statsigStrategy = new StatsigStrategy({
clientKey: 'statsig-api-key';
statsigUser: {
userId: 'my-psuedo-user-id'
};
});

await statsigStrategy.initialize();
analytics.register('statsig', statsigStrategy);
analytics.track('initialized', { test: 'test' }); // fires a tracking event to firebase and statsig
`

Event Schemas & Validation

The Analytics package is using Zod for event validation. Built-in event schemas are provided (FTM), and you can also define your own custom schemas.

Using Built-in Event Schemas

You can import the default event schemas and event names:

`ts
import { DefaultEventSchemas, EventNames } from '@curiouslearning/analytics';

// Example: Use EventNames to ensure correct event names
const eventName = EventNames.SESSION_START;

// Example: Validate event params manually (optional)
import { z } from 'zod';
const params = { / ... / };
const result = DefaultEventSchemas[eventName].safeParse(params);
if (!result.success) {
console.error(result.error);
}
`

Using StatsigStrategy with Validation

When you use the StatsigStrategy, event parameters are automatically validated against the built-in schemas (and any custom schemas you provide). Invalid events will include validation errors in the metadata.

`ts
import { StatsigStrategy, EventNames } from '@curiouslearning/analytics';

const statsigStrategy = new StatsigStrategy({
clientKey: 'your-statsig-key',
statsigUser: { userID: 'user-123' },
debug: true // enables validation error logging
});

await statsigStrategy.initialize();

// This will be validated against the SESSION_START schema
statsigStrategy.track(EventNames.SESSION_START, {
profile_number: 123,
cr_user_id: 'user123',
ftm_language: 'en_US',
version_number: '1.2.3',
json_version_number: '2.0.0',
days_since_last: 2
});
`

Common Validators (common-params)

This package provides reusable validators in common-params for required fields, such as requiredString, requiredNumber, and requiredDate. These helpers wrap Zod's types to provide consistent required field validation and clear error messages. They are used in all built-in event schemas and are recommended for custom schemas as well.

Example:
`ts
import { requiredString, requiredNumber } from '@curiouslearning/analytics/validators';

const schema = z.object({
event_date: requiredString('Event date'),
duration: requiredNumber('Duration in seconds')
});
`

- requiredString('Field Name') ensures the field is present and a string, with a custom error message.
- requiredNumber('Field Name') ensures the field is present and a number, with a custom error message.
- See common-params.ts for more helpers like requiredDate, requiredNonNegativeNumber, and requiredDateInFormat.

Defining and Using Custom Schemas

You can define your own Zod schemas and pass them to the strategy. Custom schemas will override or extend the built-in ones. For consistency and better error messages, use the provided validators from common-params:

`ts
import { z } from 'zod';
import { StatsigStrategy, EventNames } from '@curiouslearning/analytics';
import { requiredString } from '@curiouslearning/analytics/validators';

const customSchemas = {
[EventNames.SESSION_START]: z.object({
event_date: requiredString('Event date'), // uses the shared validator
custom_field: requiredString('Custom field')
})
};

const statsigStrategy = new StatsigStrategy({
clientKey: 'your-statsig-key',
statsigUser: { userID: 'user-123' },
customEventSchemas: customSchemas,
debug: true
});

await statsigStrategy.initialize();

// This will be validated against your custom SESSION_START schema
statsigStrategy.track(EventNames.SESSION_START, {
profile_number: 123,
cr_user_id: 'user123',
ftm_language: 'en_US',
version_number: '1.2.3',
json_version_number: '2.0.0',
days_since_last: 2,
custom_field: 'my-value'
});
`

Other Use Cases

A single strategy trigger

`ts
const firebaseStrategy = analytics.getRegistry('firebase');
firebaseStrategy.track('test', { test: 'test' });
`

A custom strategy

Make sure to extend the abstract strategy to making sure all necessary implementations are present.
`ts
import { AbstractAnalyticsStrategy } from '@curiouslearning/analytics';

export class CustomStrategy extends AbstractAnalyticsStrategy {
async initialize() {
// initialization
}

track(evetName: string, data: any) {
// do something with eventName and data
}
}

// ..
const customStrategy = new CustomStrategy();
await customStrategy.initialize();
analytics.register('custom', customStrategy);

analytics.track('test', { foo: 'baz' });
`

Resource Cleanup (v1.3.0+)


Both Firebase and Statsig strategies provide a dispose method to properly clean up resources when they are no longer needed. This is particularly important to prevent memory leaks and ensure proper shutdown of connections.

dispose has been added in version 1.3.0.

Firebase Strategy


`ts
const firebaseStrategy = new FirebaseStrategy({
// ... configuration ...
});

// When done with the strategy
firebaseStrategy.dispose(); // This will delete the Firebase app instance if it exists
`

Statsig Strategy


`ts
const statsigStrategy = new StatsigStrategy({
// ... configuration ...
});

// When done with the strategy
statsigStrategy.dispose(); // This will shut down the Statsig client if it exists
`

It's recommended to call dispose()` when:
- Your application is shutting down
- You're switching to a different analytics strategy
- You need to reinitialize the strategy with different configuration

Caveats

We are leaving the configuration option definition on the consuming app layer since we may want to have different buckets or measurement ids for these apps.