bigml api webinar - march 2014

19
BigML Inc API

Upload: bigml

Post on 10-May-2015

5.310 views

Category:

Technology


4 download

DESCRIPTION

Keynote from March 2014 Webinar on Building Predictive Apps with BigML's API

TRANSCRIPT

Page 1: BigML API Webinar - March 2014

BigML Inc

API

Page 2: BigML API Webinar - March 2014

BigML Inc !2

Today’s Webinar

• Speaker:

• Poul Petersen, CIO

• Moderator:

• Andrew Shikiar, VP Business Development

• Enter questions into chat box – we’ll answer some via text; others at the end of the session

• For direct follow-up, email us at [email protected]

Page 3: BigML API Webinar - March 2014

BigML Inc !3

BigML Architecturesky

wintermute

apian

https://bigml.com

https://bigml.io API Layer

Frontend Visualization Layer

Backend Computation Layer

Other Services

Page 4: BigML API Webinar - March 2014

BigML Inc

API Bindings Overview

!4

API Introduction / Demo with 1 Predictive Application Demo2

3 Programmatic ML Examples with 4

Agenda

BigMLer - a command line tool for ML5

Page 5: BigML API Webinar - March 2014

BigML Inc !5

https://bigml.io/ / /{id}?{auth}

sourcedatasetmodel

ensembleprediction

batchpredictionevaluation

andromeda

dev

dev/andromeda

• Path elements:

• /andromeda specifies the API version (optional)

• /dev specifies development mode

• if not specified, then latest API in production mode

• {id} is required for PUT and DELETE

• {auth} contains url parameters username and api_key

• api_key can be an alternative key

Page 6: BigML API Webinar - March 2014

BigML Inc !6

https://bigml.io/....{JSON} {JSON}

Operation HTTP Method Semantics

CREATE POSTCreates a new resource. Returns a JSON document including a unique identifier.

RETRIEVE GETRetrieves either a specific resource or a list of resources.

UPDATE PUT Updates a resource. Only certain fields are putable.

DELETE DELETE Deletes a resource

Page 7: BigML API Webinar - March 2014

BigML IncBigML Inc !7

Predict Color Pref?

Page 8: BigML API Webinar - March 2014

BigML IncBigML Inc !8

App Architecture

Web Server BrowserLogs

Batch Upload / Model Real-Time

request

predict

custom experience

Page 9: BigML API Webinar - March 2014

BigML IncBigML Inc !9

Log Data

user_agent colorMozilla/5.0 (Windows NT 6.1; WOW64; rv YellowMozilla/5.0 (X11; Linux i686) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1750.146 Safari/537.36 GreenMozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/32.0.1700.107 Safari/537.36 GreenMozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/32.0.1700.107 Safari/537.36 YellowMozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1750.146 Safari/537.36 RedMozilla/5.0 (Macintosh; Intel Mac OS X 10_9_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/32.0.1700.107 Safari/537.36 RedMozilla/5.0 (Macintosh; Intel Mac OS X 10_9_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1750.117 Safari/537.36 YellowMozilla/5.0 (Windows NT 6.1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1750.117 Safari/537.36 YellowMozilla/5.0 (Macintosh; Intel Mac OS X 10_9_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/33.0.1750.146 Safari/537.36 Red

Page 10: BigML API Webinar - March 2014

BigML IncBigML Inc !10

New FeaturesMozilla/5.0 (iPhone; CPU iPhone OS 7_0_6 like Mac OS X) AppleWebKit/537.51.1 (KHTML, like Gecko) Version/7.0 Mobile/11B651 Safari/9537.5

Mobile Safari

browser browser version os os version device7 iOS 7.0.6 iPhone

User-agent parser

Page 11: BigML API Webinar - March 2014

BigML IncBigML Inc !11

New Features

browser version os os version device color

Other Windows 7 Other Yellow

Chrome 33.0.1750 Linux Other Green

Chrome 32.0.1700 Windows 8 Other Green

Chrome 32.0.1700 Windows 7 Other Yellow

Chrome 33.0.1750 Windows XP Other RedChrome 32.0.1700 Mac OS X 10.9.1 Other RedChrome 33.0.1750 Mac OS X 10.9.1 Other YellowChrome 33.0.1750 Windows 7 Other YellowChrome 33.0.1750 Mac OS X 10.9.1 Other Red

Page 12: BigML API Webinar - March 2014

BigML IncBigML Inc !12

Model

Page 13: BigML API Webinar - March 2014

BigML IncBigML Inc !13

JS Predictions

. . .

Page 14: BigML API Webinar - March 2014

BigML IncBigML Inc !14

Predictions

Page 15: BigML API Webinar - March 2014

BigML IncBigML Inc !15

Gist

http://bl.ocks.org/osroca/9474489

Page 16: BigML API Webinar - March 2014

BigML Inc !16

BigML Bindings!

https://bigml.com/developers...And more:

Page 17: BigML API Webinar - March 2014

BigML Inc !17

Operation HTTP Method Binding Method

CREATE POST api.create_<resource>(from, {opts})

RETRIEVE GETapi.get_<resource>(id, {opts})

api.list_<resource>({opts})

UPDATE PUT api.update_<resource>(id, {opts})

DELETE DELETE api.delete_<resource>(id)

Binding Overview

• Where <resource> is one of: source, dataset, model, ensemble, evaluation, etc

• id is a resource identifier or resource dict

• from is a resource identifier, dict, or string depending on context

Page 18: BigML API Webinar - March 2014

BigML Inc !18

ToyBoost*

orig dataset

dataset +weight model

source +predict

batch predict

dataset +predict

*For Python Bindings Demonstration

Page 19: BigML API Webinar - March 2014

BigML Inc !19

BigMLer

•BigMLer wraps BigML’s API Python bindings

•Issue complete train/evaluation cycle in one command

•Can do cross-validation

•Remote/Local predictions or even PredictServer

•Define field types in a flat file

•Multi-Label classifications

BigMLer makes BigML even easier!