Right now we have a dataflow job and an un-related BigQuery query in the word count dag. These should be refactored to actually seem like they are part of the same workflow. For example ```python3 df_wordcount >> bq_load >> bq_top_10_query ```