Spark-zeppelin-大数据可视化分析

官网介绍

Multi-purpose Notebook

The Notebook is the place for all your needs

• Data Ingestion
• Data Discovery
• Data Analytics
• Data Visualization & Collaboration

Multiple language backend

Zeppelin interpreter concept allows any language/data-processing-backend to be plugged into Zeppelin.Currently Zeppelin supports many interpreters such as Scala(with Apache Spark), Python(with Apache Spark), SparkSQL, Hive, Markdown and Shell.

Adding new language-backend is really simple. Learn how to write a zeppelin interpreter.

Apache Spark integration

Zeppelin provides built-in Apache Spark integration. You don't need to build a separate module, plugin or library for it.

Zeppelin's Spark integration provides

• Automatic SparkContext and SQLContext injection
• Runtime jar dependency loading from local filesystem or maven repository. Learn more aboutdependency loader.
• Canceling job and displaying its progress

Data visualization

Some basic charts are already included in Zeppelin. Visualizations are not limited to SparkSQL's query, any output from any language backend can be recognized and visualized.

Pivot chart

With simple drag and drop Zeppelin aggeregates the values and display them in pivot chart. You can easily create chart with multiple aggregated values including sum, count, average, min, max.

Learn more about Zeppelin's Display system. ( text, html, table, angular )

Dynamic forms

Zeppelin can dynamically create some input forms into your notebook.

Learn more about Dynamic Forms.

Collaboration

Notebook URL can be shared among collaborators. Zeppelin can then broadcast any changes in realtime, just like the collaboration in Google docs.

Publish

Zeppelin provides an URL to display the result only, that page does not include Zeppelin's menu and buttons.This way, you can easily embed it as an iframe inside of your website.

100% Opensource

Apache Zeppelin (incubating) is Apache2 Licensed software. Please check out thesource repository andHow to contribute

Zeppelin has a very active development community.Join the Mailing list and report issues on our Issue tracker.

Undergoing Incubation

Apache Zeppelin is an effort undergoing incubation at The Apache Software Foundation (ASF), sponsored by the Incubator. Incubation is required of all newly accepted projects until a further review indicates that the infrastructure, communications, and decision making process have stabilized in a manner consistent with other successful ASF projects. While incubation status is not necessarily a reflection of the completeness or stability of the code, it does indicate that the project has yet to be fully endorsed by the ASF.

安装

From binary package

Download latest binary package from Download.

Build from source

Check instructions in README to build from source.

Configure

Configuration can be done by both environment variable(conf/zeppelin-env.sh) and java properties(conf/zeppelin-site.xml). If both defined, environment vaiable is used.

zepplin-env.sh	zepplin-site.xml	Default value	Description
ZEPPELIN_PORT	zeppelin.server.port	8080	Zeppelin server port.
ZEPPELIN_MEM	N/A	-Xmx1024m -XX:MaxPermSize=512m	JVM mem options
ZEPPELIN_INTP_MEM	N/A	ZEPPELIN_MEM	JVM mem options for interpreter process
ZEPPELIN_JAVA_OPTS	N/A		JVM Options
ZEPPELIN_ALLOWED_ORIGINS	zeppelin.server.allowed.origins	*	Allows a way to specify a ',' separated list of allowed origins for rest and websockets. i.e. http://localhost:8080
ZEPPELIN_SERVER_CONTEXT_PATH	zeppelin.server.context.path	/	Context Path of the Web Application
ZEPPELIN_SSL	zeppelin.ssl	false
ZEPPELIN_SSL_CLIENT_AUTH	zeppelin.ssl.client.auth	false
ZEPPELIN_SSL_KEYSTORE_PATH	zeppelin.ssl.keystore.path	keystore
ZEPPELIN_SSL_KEYSTORE_TYPE	zeppelin.ssl.keystore.type	JKS
ZEPPELIN_SSL_KEYSTORE_PASSWORD	zeppelin.ssl.keystore.password
ZEPPELIN_SSL_KEY_MANAGER_PASSWORD	zeppelin.ssl.key.manager.password
ZEPPELIN_SSL_TRUSTSTORE_PATH	zeppelin.ssl.truststore.path
ZEPPELIN_SSL_TRUSTSTORE_TYPE	zeppelin.ssl.truststore.type
ZEPPELIN_SSL_TRUSTSTORE_PASSWORD	zeppelin.ssl.truststore.password
ZEPPELIN_NOTEBOOK_HOMESCREEN	zeppelin.notebook.homescreen		Id of notebook to be displayed in homescreen ex) 2A94M5J1Z
ZEPPELIN_NOTEBOOK_HOMESCREEN_HIDE	zeppelin.notebook.homescreen.hide	false	hide homescreen notebook from list when this value set to "true"
ZEPPELIN_WAR_TEMPDIR	zeppelin.war.tempdir	webapps	The location of jetty temporary directory.
ZEPPELIN_NOTEBOOK_DIR	zeppelin.notebook.dir	notebook	Where notebook file is saved
ZEPPELIN_NOTEBOOK_S3_BUCKET	zeppelin.notebook.s3.bucket	zeppelin	Bucket where notebook saved
ZEPPELIN_NOTEBOOK_S3_USER	zeppelin.notebook.s3.user	user	User in bucket where notebook saved. For example bucket/user/notebook/2A94M5J1Z/note.json
ZEPPELIN_NOTEBOOK_STORAGE	zeppelin.notebook.storage	org.apache.zeppelin.notebook.repo.VFSNotebookRepo	Comma separated list of notebook storage
ZEPPELIN_INTERPRETERS	zeppelin.interpreters	org.apache.zeppelin.spark.SparkInterpreter, org.apache.zeppelin.spark.PySparkInterpreter, org.apache.zeppelin.spark.SparkSqlInterpreter, org.apache.zeppelin.spark.DepInterpreter, org.apache.zeppelin.markdown.Markdown, org.apache.zeppelin.shell.ShellInterpreter, org.apache.zeppelin.hive.HiveInterpreter ...	Comma separated interpreter configurations [Class]. First interpreter become a default
ZEPPELIN_INTERPRETER_DIR	zeppelin.interpreter.dir	interpreter	Zeppelin interpreter directory

You'll also need to configure individual interpreter. Information can be found in 'Interpreter' section in this documentation.

For example Spark.

Start/Stop

Start Zeppelin

bin/zeppelin-daemon.sh start

After successful start, visit http://localhost:8080 with your web browser.

Stop Zeppelin

bin/zeppelin-daemon.sh stop

实践例子：

Zeppelin Tutorial

We will assume you have Zeppelin installed already. If that's not the case, seeInstall.

Zeppelin's current main backend processing engine is Apache Spark. If you're new to the system, you might want to start by getting an idea of how it processes data to get the most out of Zeppelin.

Tutorial with Local File

Data Refine

Before you start Zeppelin tutorial, you will need to download bank.zip.

First, to transform data from csv format into RDD of Bank objects, run following script. This will also remove header usingfilter function.

val bankText = sc.textFile("yourPath/bank/bank-full.csv")

case class Bank(age:Integer, job:String, marital : String, education : String, balance : Integer)

// split each line, filter out header (starts with "age"), and map it into Bank case class 
val bank = bankText.map(s=>s.split(";")).filter(s=>s(0)!="\"age\"").map(
    s=>Bank(s(0).toInt, 
            s(1).replaceAll("\"", ""),
            s(2).replaceAll("\"", ""),
            s(3).replaceAll("\"", ""),
            s(5).replaceAll("\"", "").toInt
        )
)

// convert to DataFrame and create temporal table
bank.toDF().registerTempTable("bank")

Data Retrieval

Suppose we want to see age distribution from bank. To do this, run:

%sql select age, count(1) from bank where age < 30 group by age order by age

You can make input box for setting age condition by replacing 30 with${maxAge=30}.

%sql select age, count(1) from bank where age < ${maxAge=30} group by age order by age

Now we want to see age distribution with certain marital status and add combo box to select marital status. Run:

%sql select age, count(1) from bank where marital="${marital=single,single|divorced|married}" group by age order by age

Tutorial with Streaming Data

Data Refine

Since this tutorial is based on Twitter's sample tweet stream, you must configure authentication with a Twitter account. To do this, take a look atTwitter Credential Setup. After you get API keys, you should fill out credential related values(apiKey,apiSecret, accessToken, accessTokenSecret) with your API keys on following script.

This will create a RDD of Tweet objects and register these stream data as a table:

import org.apache.spark.streaming._
import org.apache.spark.streaming.twitter._
import org.apache.spark.storage.StorageLevel
import scala.io.Source
import scala.collection.mutable.HashMap
import java.io.File
import org.apache.log4j.Logger
import org.apache.log4j.Level
import sys.process.stringSeqToProcess

/** Configures the Oauth Credentials for accessing Twitter */
def configureTwitterCredentials(apiKey: String, apiSecret: String, accessToken: String, accessTokenSecret: String) {
  val configs = new HashMap[String, String] ++= Seq(
    "apiKey" -> apiKey, "apiSecret" -> apiSecret, "accessToken" -> accessToken, "accessTokenSecret" -> accessTokenSecret)
  println("Configuring Twitter OAuth")
  configs.foreach{ case(key, value) =>
    if (value.trim.isEmpty) {
      throw new Exception("Error setting authentication - value for " + key + " not set")
    }
    val fullKey = "twitter4j.oauth." + key.replace("api", "consumer")
    System.setProperty(fullKey, value.trim)
    println("\tProperty " + fullKey + " set as [" + value.trim + "]")
  }
  println()
}

// Configure Twitter credentials
val apiKey = "xxxxxxxxxxxxxxxxxxxxxxxxx"
val apiSecret = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
val accessToken = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
val accessTokenSecret = "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
configureTwitterCredentials(apiKey, apiSecret, accessToken, accessTokenSecret)

import org.apache.spark.streaming.twitter._
val ssc = new StreamingContext(sc, Seconds(2))
val tweets = TwitterUtils.createStream(ssc, None)
val twt = tweets.window(Seconds(60))

case class Tweet(createdAt:Long, text:String)
twt.map(status=>
  Tweet(status.getCreatedAt().getTime()/1000, status.getText())
).foreachRDD(rdd=>
  // Below line works only in spark 1.3.0.
  // For spark 1.1.x and spark 1.2.x,
  // use rdd.registerTempTable("tweets") instead.
  rdd.toDF().registerAsTable("tweets")
)

twt.print

ssc.start()

Data Retrieval

For each following script, every time you click run button you will see different result since it is based on real-time data.

Let's begin by extracting maximum 10 tweets which contain the word "girl".

%sql select * from tweets where text like '%girl%' limit 10

This time suppose we want to see how many tweets have been created per sec during last 60 sec. To do this, run:

%sql select createdAt, count(1) from tweets group by createdAt order by createdAt

You can make user-defined function and use it in Spark SQL. Let's try it by making function namedsentiment. This function will return one of the three attitudes(positive, negative, neutral) towards the parameter.

def sentiment(s:String) : String = {
    val positive = Array("like", "love", "good", "great", "happy", "cool", "the", "one", "that")
    val negative = Array("hate", "bad", "stupid", "is")

    var st = 0;

    val words = s.split(" ")    
    positive.foreach(p =>
        words.foreach(w =>
            if(p==w) st = st+1
        )
    )

    negative.foreach(p=>
        words.foreach(w=>
            if(p==w) st = st-1
        )
    )
    if(st>0)
        "positivie"
    else if(st<0)
        "negative"
    else
        "neutral"
}

// Below line works only in spark 1.3.0.
// For spark 1.1.x and spark 1.2.x,
// use sqlc.registerFunction("sentiment", sentiment _) instead.
sqlc.udf.register("sentiment", sentiment _)

To check how people think about girls using sentiment function we've made above, run this:

%sql select sentiment(text), count(1) from tweets where text like '%girl%' group by sentiment(text)