第十四章

MongoDB與Python數(shù)據(jù)分析單擊此處添加副標(biāo)題演講者：環(huán)境準(zhǔn)備pipinstallpandaspipinstallnumpypipinstallpyechartspipinstalljieba環(huán)境準(zhǔn)備認(rèn)識(shí)pyechartsfrompyecharts.chartsimportBar

bar=(

Bar()

.add_xaxis(["外套","羊毛衫","短袖","長(zhǎng)褲","靴子","連衣裙"])

.add_yaxis("商家A",[25,15,35,12,75,85]))bar.render('html/pyechartsdemo.html')電商數(shù)據(jù)分析讀取數(shù)據(jù)電商數(shù)據(jù)分析處理數(shù)據(jù)刪除重復(fù)值：user.drop_duplicates(keep='last',inplace=True)將time轉(zhuǎn)換為datetime格式：user['time']=pd.to_datetime(user['time'])提取出日期和時(shí)間：user['dates']=user.time.dt.dateuser['month']=user.dates.values.astype('datetime64[M]')user['hours']=user.time.dt.hour轉(zhuǎn)換數(shù)據(jù)類型：user['behavior_type']=user['behavior_type'].apply(str)

#行為類型user['user_id']=user['user_id'].apply(str)user['item_id']=user['item_id'].apply(str)電商數(shù)據(jù)分析數(shù)據(jù)分析

：統(tǒng)計(jì)每日訪問(wèn)量和獨(dú)立訪客量數(shù)據(jù)電商數(shù)據(jù)分析數(shù)據(jù)分析

：按天統(tǒng)計(jì)訪客量和獨(dú)立訪客量的分析差異電商數(shù)據(jù)分析數(shù)據(jù)分析

：不同時(shí)期用戶行為分析電商數(shù)據(jù)分析數(shù)據(jù)分析

：活動(dòng)期間不同時(shí)段的用戶行為分析電商數(shù)據(jù)分析數(shù)據(jù)分析

：日常期間不同時(shí)段的用戶行為分析電商數(shù)據(jù)分析數(shù)據(jù)分析

：不同時(shí)段購(gòu)買率電商數(shù)據(jù)分析數(shù)據(jù)分析

：活動(dòng)期間轉(zhuǎn)化率分析電商數(shù)據(jù)分析數(shù)據(jù)分析

：日常期間轉(zhuǎn)化率分析端午節(jié)粽子數(shù)據(jù)分析讀取數(shù)據(jù)mon_client=pymongo.MongoClient(host='Localhost',port=27017)mon_db=mon_client['duanwu']mon_col=mon_db['zongzi']res=mon_col.find().limit(10)df=pd.DataFrame(list(res))print(df)提取付款人數(shù)單位（萬(wàn)）：df['unit']=[''.join(re.findall(r'(萬(wàn))',i))foriindf['付款人數(shù)']]#提取單位（萬(wàn)）df['unit']=df['unit'].apply(lambdax:10000ifx=='萬(wàn)'else1)計(jì)算銷量：df['銷量']=df['num']*df['unit']刪除無(wú)發(fā)貨地址的商品，并提取省份：df=df[df['發(fā)貨地址'].notna()]df['省份']=df['發(fā)貨地址'].str.split('').apply(lambdax:x[0])刪除多余的列：df.drop(['付款人數(shù)','發(fā)貨地址','num','unit'],axis=1,inplace=True)重置索引：df=df.reset_index(drop=True)端午節(jié)粽子數(shù)據(jù)分析處理數(shù)據(jù)去除重復(fù)值df.drop_duplicates(inplace=True)處理購(gòu)買人數(shù)為空的記錄修改為“0人付款”，方便后面做統(tǒng)計(jì)分析：df['付款人數(shù)']=df['付款人數(shù)'].replace(np.nan,'0人付款')提取付款人數(shù)的數(shù)值并轉(zhuǎn)化數(shù)據(jù)類型：df['num']=[re.findall(r'(\d+\.{0,1}\d*)',i)[0]foriindf['付款人數(shù)']]#提取數(shù)值df['num']=df['num'].astype('float')#轉(zhuǎn)化數(shù)值型添加停用詞:stop_words.extend(['logo','10','100','200g','100g','140g','130g'])評(píng)論字段分詞處理word_num=jieba.analyse.extract_tags(txt,topK=100,withWeight=TrueallowPOS=())去停用詞word_num_selected=[]?foriinword_num:

ifi[0]notinstop_words:

word_num_selected.append(i)?key_words=pd.DataFrame(word_num_selected,columns=['words','num'])端午節(jié)粽子數(shù)據(jù)分析數(shù)據(jù)分析：數(shù)據(jù)準(zhǔn)備-分詞importjieba.analysetxt=df['商品名'].str.cat(sep='。')添加關(guān)鍵詞：jieba.add_word('粽子',999,'五芳齋')讀入停用詞表：stop_words=[]withopen('停用詞表/哈工大停用詞表.txt','rb')asf:

lines=f.readlines()

forlineinlines:

stop_words.append(line.strip())端午節(jié)粽子數(shù)據(jù)分析數(shù)據(jù)分析：端午節(jié)粽子數(shù)據(jù)分析數(shù)據(jù)分析：店鋪銷量排名前10端午節(jié)粽子數(shù)據(jù)分析數(shù)據(jù)分析：不同