github地址
us_baby_names项目地址
tv_inspired_baby_names项目地址
运行
- 使用环境
python 2.7(运行过程中缺少包自行安装)
jupyter 1.0.0 安装使用 - 运行代码
##cd进.ipynb 所在目录
cd python
## 开启 jupyter
jupyter notebook
点击对应文件(注意要先把数据集下载到对应目录)
然后点击运行,等待一段时间就会出现结果。结果如下图。
源码分析
- 源数据数据格式
源数据就是id与分年份与性别的统计数据(一份有大洲数据)。 - 代码分析
2.1 us_baby_names
## 读取csv数据到data中
data = pd.read_csv('./dataset/NationalNames.csv')
##展示data数据头部5条信息
data.head()
## 查看数据详情
data.info()
# 将“姓名”和其数量进行统计,并放入“字典”变量中
names_dict = dict()
# 由于数据量多,耗时较长
for index, row in data.iterrows():
# 按行遍历数据 如果没有该名字,则新建字典index否则加一
if row['Name'] not in names_dict:
names_dict[row['Name']] = row['Count']
else:
names_dict[row['Name']] += row['Count']
# 使用Counter计数器的most_common函数进行统计。
# 返回最常用的元素及其计数的列表
top_10 = Counter(names_dict).most_common(10)
print '全美流行的婴儿名字top10:'
for pair in top_10:
print '姓名:%s -> 数量:%i' %(pair[0], pair[1])
def average_length_data_transform():
"""
统计每年男性、女性姓名的平均长度
"""
# 按行遍历数据
years = []
# 女性姓名
female_average_length = []
female_average_name_length = dict()
# 男性姓名
male_average_length = []
male_average_name_length = dict()
for index, row in data.iterrows():
# 按行遍历数据集
if row['Gender'] == 'F':
# 女性
# 获取当前年份
curr_year = row['Year']
# 当前记录姓名长度
curr_name_length = len(row['Name'])
if curr_year not in female_average_name_length:
# 如果当前年份不在“字典”中,放入字典中
# 比如: 1880年第一条女性记录:1 Mary 1880 F 7065,将做如下处理
# {1880: [4, 1]}
female_average_name_length[curr_year] = [curr_name_length, 1]
else:
# 比如: 1880年第二条女性记录: 2 Anna 1880 F 2604,将做如下处理
# {1880: [4+4, 1+1]} -> {1880: [8, 2]}
female_average_name_length[curr_year][0] += curr_name_length
female_average_name_length[curr_year][1] += 1
else:
# 男性
# 处理过程同上
curr_year = row['Year']
curr_name_length = len(row['Name'])
if curr_year not in male_average_name_length:
male_average_name_length[curr_year] = [curr_name_length, 1]
else:
male_average_name_length[curr_year][0] += curr_name_length
male_average_name_length[curr_year][1] += 1
# 遍历处理后的字典
# 女性
for key, value in female_average_name_length.items():
years.append(key)
female_average_length.append(float(value[0]) / value[1])
# 男性
for key, value in male_average_name_length.items():
years.append(key)
male_average_length.append(float(value[0]) / value[1])
return (female_average_length, female_average_name_length, male_average_length, male_average_name_length)
# years = [1880, 1881, ..., 2014]
years = range(1880, 2015)
# 使用matplotlib进行数据可视化
f, ax = plt.subplots(figsize=(10, 6))
# 设置x轴数值范围
ax.set_xlim([1880, 2014])
# years为x轴,average_length为y轴
# 女性曲线为红色,男性为蓝色
plt.plot(years, female_average_length, label='Average length of female names', color='r')
plt.plot(years, male_average_length, label='Average length of male names', color='b')
# 设置x,y轴标签
ax.set_ylabel('Length of name')
ax.set_xlabel('Year')
# 设置图像名称
ax.set_title('Average length of names')
# 设置图例
legend = plt.legend(loc='best', frameon=True, borderpad=1, borderaxespad=1)
唯一姓名分析
top_in_each_year = dict()
years = range(1880, 2015)
for each_year in years:
#获取每一年的数据(为啥这样用???)
each_year_data = data[data['Year'] == each_year]
top_in_each_year[each_year] = dict()
#获取每一年的某一姓名的数量
for index, row in each_year_data.iterrows():
# 为什么不是+=(可能数据是不重复的??)
top_in_each_year[each_year][row['Name']] = row['Count']
all_sum = []
top_25_sum = []
#按照年份遍历数组
for year, names_in_year in top_in_each_year.items():
#将当前年份总人数添加到all_sum中
all_sum.append(sum(Counter(names_in_year).values()))
top_25 = Counter(names_in_year).most_common(25)
sum_temp = 0
for pair in top_25:
sum_temp += pair[1]
# 将当前出现最频繁的25个名字总数相加和加入到列表
top_25_sum.append(sum_temp)
#可视化
......
2.2 tv_inspired_baby_names
def plotname(name, gender):
"""
对姓名及性别进行可视化
"""
# 找到符合条件的数据
data_named = data[(data.Name==name) & (data.Gender==gender)]
plt.figure(figsize=(10,6))
# 对数据的年份,数量进行可视化
plt.plot(data_named.Year,data_named.Count,'g-')
plt.title('%s name variation over time'%name)
plt.ylabel('counts')
plt.xticks(data_named.Year,rotation='vertical')
# 设置x轴数值范围
plt.xlim([1985,2015])
plt.show()
MMP 又挂了,图片上传不了...
总结
这份代码其实就是python以及可视化,数据处理等包的简单使用,感觉并没有很大的难度,主要是环境搭建和一些数据处理函数的使用。