In chriszheng2016/ZStockModels: ZStockModels - Stock factors Models for China Stock Market Investment

研究各股票的利润情况

library(tidyverse)
library(lubridate)
library(stringr)
library(shiny)
library(ZStockModels)

# Set global options for knitr 
knitr::opts_chunk$set(
comment = "#>",
echo = FALSE,
warning = FALSE,
message = FALSE,
collapse = TRUE
)

options(tibble.print_min = 10L, tibble.print_max = 20L)

Data Preparation(数据准备)

Data tables in database(数据库所包含数据表)

# Open gta stock databse
stock_db <- stock_db(gta_db, "GTA_SQLData")
open_stock_db(stock_db)

# Initate the stock database
invisible(init_stock_db(stock_db))
stock_name_list  <- stock_name_list(stock_db)
stock_field_list <- stock_field_list(stock_db)

# Get names of all stock table from database   
 db_tables <- list_stock_tables(stock_db)
 knitr::kable(
  x = db_tables,
  caption = "Tables in GTA Stock Database",
  align = "l"
 )

共有r length(db_tables)个数据表

Summary of Stocks&Industries in Universe(全部股票及行业信息)

全部股票与行业信息如下：

# Get company info 
trd_company.df <- get_table_dataset(stock_db, table_name = "TRD_Co_公司基本情况")
trd_company.tib <- tibble::as.tibble(trd_company.df)

# all stocks in the investment universe
universe_stocks <- trd_company.tib %>%
    select(stkcd, stkname= stknme, indcd = nnindcd, indname = nnindnme) %>%
    arrange(indcd, stkcd)


# all industries in investment universe
universe_industries <- universe_stocks %>%
  group_by(indcd, indname) %>%
  summarise(stknum = n()) %>%
  arrange(indcd, desc(stknum))

# Output the results
universe_industries
universe_stocks

共r nrow(universe_industries)个行业, r nrow(universe_stocks)只股票

Set Params of Study(设置研究参数)

组合股票明细：

# Set params of study

table_indicator <- "FR_T5_盈利能力"

# Set stocks as portfolio
stock_stkcds_list <- c(600066,000550, 600031, 000157,000651, 000333)
portfolio_stocks <- universe_stocks %>%
    filter(stkcd %in% stock_stkcds_list)

#Set focus indicator of portfolio
focus_fields_code <- c("f053301c", "f050504c", "f052401c")
focus_fields_name <- code2name(stock_field_list,focus_fields_code)
focus_fields.tib  <- tibble(code = focus_fields_code, name = focus_fields_name)

the_indicator <- focus_fields_code[2]
the_indicator_name <- code2name(stock_field_list,the_indicator)

# Set top/last n industries for show
top_last_n_industies <- 5

#Output the results
portfolio_stocks

focus_fields.tib

研究指标: r str_c(the_indicator, "/",the_indicator_name)

Load Inicators of Porfolio(加载组合指标数据)

加载组合指标数据如下：

# Load indicators of all stocks
universe_data.df <- as.tibble(get_table_dataset(stock_db, table_name = table_indicator))
universe_data.tib <- as.tibble(universe_data.df)

# get indicators of all stocks
universe_data <- universe_data.tib %>%
    left_join(trd_company.tib, by  = "stkcd") %>%
    mutate(periodtype = ifelse(lubridate::month(accper) == 12, "annual","quarter")) %>%
    filter(typrep == 'A') %>%
    select(accper, periodtype, stkcd, stkname = stknme, 
           indcd = indcd.x, indname = nnindnme,
           contains("0"))

# get indicators of portfolio
portfolio_data <- universe_data %>%
      dplyr::filter( stkcd %in% stock_stkcds_list)


(portfolio_data)

组合指标字段信息如下：

# Get all fields of portfolio 
stock_field_code_list <- colnames(portfolio_data)
stock_field_code_list <- stringr::str_subset(stock_field_code_list, "\\d")
stock_field_name_list <- code2name(stock_field_list, stock_field_code_list)
portfolio_fields <- tibble::tibble(code = stock_field_code_list, name = stock_field_name_list)

# get indicators data for allstock
universe_indicators <- universe_data %>%
  select(accper:indname, the_indicator) 

# get indictors data for portfolio
portfolio_indicators <- portfolio_data %>%
  select(accper:indname, the_indicator) 

# Output the results
portfolio_fields

portfolio_indicators

研究指标: r str_c(the_indicator, "/",the_indicator_name)

Data Exploration(数据探索)

Industry Study (各行业相关指标研究)

1.各行业指标数据统计分布

各行业各指标中值分布如下：

# use median of each peroid as indicator for industry
industry_indicator_peroid<- universe_indicators %>%
    group_by(indname, accper) %>%
    summarise_at(the_indicator, .funs = funs(mean, median, sd, max, min)) %>%
    mutate(periodtype = ifelse(lubridate::month(accper) == 12, 
                                "annual","quarter")) %>%
    select(indname, accper, periodtype, !!the_indicator:=median, everything()) %>%
    arrange(indname, accper) %>%
    ungroup()

# Indicator summary of each industry
industry_indicator_summary <- industry_indicator_peroid %>%
    group_by(indname) %>%
    summarise_at(the_indicator, .funs = funs(mean, median, sd, max, min)) %>%
    filter(!is.na(indname)) %>%
    arrange(desc(mean))

# Output the results
industry_indicator_summary


# top and last industries
sort_var <- "mean"
top_industries <- industry_indicator_summary %>%
     top_n(top_last_n_industies, get(sort_var)) %>%
     select(indname) 
top_industries <- top_industries$indname

last_industries <- industry_indicator_summary %>%
     top_n(-top_last_n_industies, get(sort_var)) %>%
     select(indname)
last_industries <- last_industries$indname



# Top 10 industries of Indicator
industry_indicator_summary %>%
  filter(indname %in% top_industries) %>%
  ggplot(aes(x = reorder(indname, get(sort_var) ), y = get(sort_var))) +
    geom_col() +
    labs(
      title = str_c("Top Industries of Indicator"),
      subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
      x = "Industry",
      y = "Median of Indicator"
    ) +
    geom_label(aes(label = format(get(sort_var), digits = 2, nsmall = 2)), hjust = TRUE) +
    coord_flip()

# last 10 industries of Indicator
industry_indicator_summary %>%
  filter(indname %in% last_industries) %>%
  ggplot(aes(x = reorder(indname, get(sort_var)), y = get(sort_var))) +
    geom_col() +
    labs(
        title = str_c("Last Industries of Indicator"),
        subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
        x = "Industry",
        y = "Median of Indicator"
      ) +
    geom_label(aes(label = format(get(sort_var), digits = 2, nsmall = 2)), hjust = FALSE) +
    coord_flip()

# Plot histogram of distiubtion of top industry
industry_indicator_peroid %>%
    filter(indname %in% top_industries) %>%
    ggplot(aes(x = get(the_indicator))) +
      geom_histogram(binwidth = 0.01) +
      geom_density() +
      labs(
      title = "Distribution Summarry of Top Industries",
      subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
      x = NULL,
      y = NULL
      ) +
      coord_cartesian(xlim = c(-1, 1)) +
      facet_wrap(~indname)

# Plot histogram of distiubtion of last industry
industry_indicator_peroid %>%
    filter(indname %in% last_industries) %>%
    ggplot(aes(x = get(the_indicator))) +
      geom_histogram(binwidth = 0.01) +
      geom_density() +
      labs(
      title = "Distribution Summarry of of Last Industries ",
      subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
      x = NULL,
      y = NULL
      ) +
      coord_cartesian(xlim = c(-1, 1)) +
      facet_wrap(~indname)


# industry_indicator_peroid %>%
#    filter(indname %in% top_industries) %>%
#    ggplot(aes(x = reorder(indname, median), y = median)) +
#    geom_boxplot() +
#    coord_flip()

2.各行业指标数据时间趋势

# plot data for top industries 

# plot series of annual data
industry_indicator_peroid %>%
    filter(indname %in% top_industries) %>%
    filter(periodtype == "annual") %>%
    ggplot(aes(x = accper, y = get(the_indicator))) +
       geom_col() + 
       labs(
         title = "Annual Trends of Top Industries",
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
       coord_cartesian(ylim = c(-1, 1)) +
       facet_wrap(~indname) 



# plot series of quarter data
industry_indicator_peroid %>%
    filter(indname %in% top_industries) %>%
    ggplot(aes(x = accper, y = get(the_indicator))) +
       geom_col() + 
       labs(
         title = "Quarterly Trends of Top Industries", 
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
       coord_cartesian(ylim = c(-1, 1)) +
       facet_wrap(~indname) 

# plot data for last industries

# plot series of annual data
industry_indicator_peroid %>%
    filter(indname %in% last_industries) %>%
    filter(periodtype == "annual") %>%
    ggplot(aes(x = accper, y = get(the_indicator))) +
       geom_col() + 
       labs(
         title = "Annual Trends of Last Industries",
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
       coord_cartesian(ylim = c(-1, 1)) +
       facet_wrap(~indname) 



# plot series of quarter data
industry_indicator_peroid %>%
    filter(indname %in% last_industries) %>%
    ggplot(aes(x = accper, y = get(the_indicator))) +
       geom_col() + 
       labs(
         title = "Quarterly Trends of Last Industries", 
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
       coord_cartesian(ylim = c(-1, 1)) +
       facet_wrap(~indname) 

Portfoilo Study(投资组合相关指标研究)

1.各股指标数据统计分布

# summary distribution of each stock
portfolio_indicators_summary <- portfolio_indicators %>%
  group_by(stkname) %>%
  summarise(
            mean = mean(get(the_indicator), na.rm = TRUE),
            std  = sqrt(var(get(the_indicator), na.rm = TRUE)),
            max = max(get(the_indicator), na.rm = TRUE),
            Q3 = quantile(get(the_indicator),probs = c(0.75),na.rm = TRUE),
            median = median(get(the_indicator), na.rm = TRUE),
            Q1 = quantile(get(the_indicator),probs = c(0.25),na.rm = TRUE),
            min = min(get(the_indicator), na.rm = TRUE)
            )

# portfolio_indicators_summary

 knitr::kable(
  x = portfolio_indicators_summary,
  caption = "Portfolio Indicator Summary",
  align = "l"
 )

# Plot boxplot of distribution of each stock
portfolio_indicators %>%
  ggplot(aes(x = reorder(stkname, get(the_indicator), FUN = median), 
             y = get(the_indicator), 
             colour = indname)) +
     geom_boxplot() +
     labs(
         title = "Box Summarry ",
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
     theme(legend.position = "bottom") +
     coord_flip()

# Plot histogram of distiubtion of each stock 
portfolio_indicators %>%
   ggplot(aes(x = get(the_indicator))) +
      geom_histogram(binwidth = 0.01) +
      geom_density() +
      labs(
         title = "Distribution Summarry ",
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
      facet_wrap(~stkname)

2.各股指标数据时间趋势

# plot series of annual data
portfolio_indicators %>%
    filter(periodtype == "annual") %>%
    ggplot(aes(x = accper, y = get(the_indicator))) +
       geom_col() + 
       labs(
         title = "Annual Trends",
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
       facet_wrap(~stkname) 



# plot series of quarter data
portfolio_indicators %>%
    ggplot(aes(x = accper, y = get(the_indicator))) +
       geom_col() + 
       labs(
         title = "Quarterly Trends", 
         subtitle = str_c("(", the_indicator, "/", the_indicator_name, ")"),
         x = NULL,
         y = NULL
         ) +
       facet_wrap(~stkname) 

# Close gta stock databse
close_stock_db(stock_db)

chriszheng2016/ZStockModels documentation built on May 11, 2019, 6:26 p.m.